From bparees at redhat.com Wed Jun 1 02:58:37 2016 From: bparees at redhat.com (Ben Parees) Date: Tue, 31 May 2016 22:58:37 -0400 Subject: [kontinuity-dev-public] "Build slaves" in Jenkins pipelines In-Reply-To: <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> References: <1464704380.9457.63.camel@redhat.com> <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> Message-ID: On Tue, May 31, 2016 at 10:27 AM, James Strachan wrote: > I?m a huge fan of 2); but currently that requires a pod for the JNLP slave > and another pod for the docker image; which on premise or on Dedicated > doesn?t really matter. But on OpenShift Online that is gonna be painful (3 > pods just for builds!); so option 1) using one or two curated pools of > slaves of custom pod/mages might be simpler for Online. e.g. one slave pool > for maven/java and one for nodejs? > ?yeah so i've been leaning towards (1) due to it needing fewer pods, but it does mean we have to provide some slave images, and/or an easy way to create new slave images (which we actually already have here: https://github.com/openshift/origin/blob/master/examples/jenkins/master-slave/jenkins-slave-template.json), so certainly (2) feels like a better long term solution. Providing maven/java and nodejs slave images out of the box seems likely. ? > > > On 31 May 2016, at 15:19, Tomas Nozicka wrote: > > Hi, > > how do we want to run our Jenkins "build slaves" in our pipelines? > > Options: > 1) kubernetes plugin [1] > = Example = > node ('java8-maven') { > sh 'mvn package' > } > = EOF = > - tightly coupled to Jenkins instance (requires predefined pools of > slaves which you can choose from) > - build environment is defined by admin not pipeline (Jenkinsfile) > - not generic > > 2) kubernetes-workflow plugin [2] > = Example = > kubernetes.pod('buildpod').withImage('java8-maven').inside{ > sh 'mvn package' > } > = EOF = > + loosely coupled to Jenkins instance > + pipeline (Jenkinsfile) specifies build environment > + generic (pipelines can specify custom images) > > I am for option 2 because of the loose coupling to Jenkins and ability > to specify (custom) build images as part of Jenkinsfile. > > And AFAIK catapult project is partially about taking user pipelines and > setting those up with OpenShift(+Jenkins), I just don't see it working > with the tight coupling and option 1. > > If we decide to go with option 2 we should probably integrate > kubernetes-workflow plugin into jenkins-1-centos7:dev image. > > What do you think? > > Thanks, > Tomas > > [1] - https://wiki.jenkins-ci.org/display/JENKINS/Kubernetes+Plugin > [2] - https://github.com/fabric8io/kubernetes-workflow > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > James > ------- > Red Hat > > Twitter: @jstrachan > Email: jstracha at redhat.com > Blog: https://medium.com/@jstrachan/ > > fabric8: http://fabric8.io/ > open source microservices platform > > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Wed Jun 1 12:37:51 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Wed, 01 Jun 2016 14:37:51 +0200 Subject: [kontinuity-dev-public] "Build slaves" in Jenkins pipelines In-Reply-To: <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> References: <1464704380.9457.63.camel@redhat.com> <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> Message-ID: <1464784671.3614.14.camel@redhat.com> Inline. On ?t, 2016-05-31 at 15:27 +0100, James Strachan wrote: > I?m a huge fan of 2); but currently that requires a pod for the JNLP > slave and another pod for the docker image; which on premise or on James, why does it need 2 pods? Could you elaborate on that? I am thinking more in terms of 1 pod with two containers (JNLP and builder image) or one builder image with JNLP embedded. Is this something that can be fixed in future? Thanks, Tomas > Dedicated doesn?t really matter. But on OpenShift Online that is > gonna be painful (3 pods just for builds!); so option 1) using one or > two curated pools of slaves of custom pod/mages might be simpler for > Online. e.g. one slave pool for maven/java and one for nodejs? > > > On 31 May 2016, at 15:19, Tomas Nozicka > > wrote: > > > > Hi, > > > > how do we want to run our Jenkins "build slaves" in our pipelines? > > > > Options: > > 1) kubernetes plugin [1] > > = Example = > > node ('java8-maven') { > > ? sh 'mvn package' > > } > > = EOF = > > ?- tightly coupled to Jenkins instance (requires predefined pools > > of > > slaves which you can choose from) > > ?- build environment is defined by admin not pipeline (Jenkinsfile) > > ?- not generic > > > > 2) kubernetes-workflow plugin [2] > > = Example = > > kubernetes.pod('buildpod').withImage('java8-maven').inside{ > > ? sh 'mvn package' > > } > > =?EOF = > > ?+ loosely coupled to Jenkins instance > > ?+ pipeline (Jenkinsfile) specifies build environment > > ?+ generic (pipelines can specify custom?images) > > > > I am for option 2 because of the loose coupling to Jenkins and > > ability > > to specify (custom) build images as part of Jenkinsfile. > > > > And AFAIK catapult project is partially about taking user pipelines > > and > > setting those up with OpenShift(+Jenkins), I just don't see it > > working > > with the tight coupling and option 1. > > > > If we decide to go with option 2 we should probably integrate > > kubernetes-workflow plugin into jenkins-1-centos7:dev image. > > > > What do you think? > > > > Thanks, > > Tomas > > > > [1] -?https://wiki.jenkins-ci.org/display/JENKINS/Kubernetes+Plugin > > [2] -?https://github.com/fabric8io/kubernetes-workflow > > > > _______________________________________________ > > kontinuity-dev-public mailing list > > kontinuity-dev-public at redhat.com > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > James > ------- > Red Hat > > Twitter: @jstrachan > Email: jstracha at redhat.com > Blog: https://medium.com/@jstrachan/ > > fabric8: http://fabric8.io/ > open source microservices platform > From jstracha at redhat.com Wed Jun 1 12:46:55 2016 From: jstracha at redhat.com (James Strachan) Date: Wed, 1 Jun 2016 13:46:55 +0100 Subject: [kontinuity-dev-public] "Build slaves" in Jenkins pipelines In-Reply-To: <1464784671.3614.14.camel@redhat.com> References: <1464704380.9457.63.camel@redhat.com> <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> <1464784671.3614.14.camel@redhat.com> Message-ID: > On 1 Jun 2016, at 13:37, Tomas Nozicka wrote: > > Inline. > > On ?t, 2016-05-31 at 15:27 +0100, James Strachan wrote: >> I?m a huge fan of 2); but currently that requires a pod for the JNLP >> slave and another pod for the docker image; which on premise or on > James, why does it need 2 pods? Could you elaborate on that? one for JNLP and one for the docker images listed in the kubernetes.pod.withImage statement. > I am thinking more in terms of 1 pod with two containers (JNLP and builder image) or one builder image with JNLP embedded. > > Is this something that can be fixed in future? Any contributions most welcome! https://github.com/fabric8io/kubernetes-workflow/issues/40 > > Thanks, > Tomas > >> Dedicated doesn?t really matter. But on OpenShift Online that is >> gonna be painful (3 pods just for builds!); so option 1) using one or >> two curated pools of slaves of custom pod/mages might be simpler for >> Online. e.g. one slave pool for maven/java and one for nodejs? >> >>> On 31 May 2016, at 15:19, Tomas Nozicka >>> wrote: >>> >>> Hi, >>> >>> how do we want to run our Jenkins "build slaves" in our pipelines? >>> >>> Options: >>> 1) kubernetes plugin [1] >>> = Example = >>> node ('java8-maven') { >>> sh 'mvn package' >>> } >>> = EOF = >>> - tightly coupled to Jenkins instance (requires predefined pools >>> of >>> slaves which you can choose from) >>> - build environment is defined by admin not pipeline (Jenkinsfile) >>> - not generic >>> >>> 2) kubernetes-workflow plugin [2] >>> = Example = >>> kubernetes.pod('buildpod').withImage('java8-maven').inside{ >>> sh 'mvn package' >>> } >>> = EOF = >>> + loosely coupled to Jenkins instance >>> + pipeline (Jenkinsfile) specifies build environment >>> + generic (pipelines can specify custom images) >>> >>> I am for option 2 because of the loose coupling to Jenkins and >>> ability >>> to specify (custom) build images as part of Jenkinsfile. >>> >>> And AFAIK catapult project is partially about taking user pipelines >>> and >>> setting those up with OpenShift(+Jenkins), I just don't see it >>> working >>> with the tight coupling and option 1. >>> >>> If we decide to go with option 2 we should probably integrate >>> kubernetes-workflow plugin into jenkins-1-centos7:dev image. >>> >>> What do you think? >>> >>> Thanks, >>> Tomas >>> >>> [1] - https://wiki.jenkins-ci.org/display/JENKINS/Kubernetes+Plugin >>> [2] - https://github.com/fabric8io/kubernetes-workflow >>> >>> _______________________________________________ >>> kontinuity-dev-public mailing list >>> kontinuity-dev-public at redhat.com >>> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> >> James >> ------- >> Red Hat >> >> Twitter: @jstrachan >> Email: jstracha at redhat.com >> Blog: https://medium.com/@jstrachan/ >> >> fabric8: http://fabric8.io/ >> open source microservices platform >> James ------- Red Hat Twitter: @jstrachan Email: jstracha at redhat.com Blog: https://medium.com/@jstrachan/ fabric8: http://fabric8.io/ open source microservices platform -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Wed Jun 1 14:08:04 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Wed, 01 Jun 2016 16:08:04 +0200 Subject: [kontinuity-dev-public] "Build slaves" in Jenkins pipelines In-Reply-To: References: <1464704380.9457.63.camel@redhat.com> <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> Message-ID: <1464790084.3614.50.camel@redhat.com> Inline. On ?t, 2016-05-31 at 22:58 -0400, Ben Parees wrote: > > > On Tue, May 31, 2016 at 10:27 AM, James Strachan > wrote: > > I?m a huge fan of 2); but currently that requires a pod for the > > JNLP slave and another pod for the docker image; which on premise > > or on Dedicated doesn?t really matter. But on OpenShift Online that > > is gonna be painful (3 pods just for builds!); so option 1) using > > one or two curated pools of slaves of custom pod/mages might be > > simpler for Online. e.g. one slave pool for maven/java and one for > > nodejs? > > > yeah so i've been leaning towards (1) due to it needing fewer pods, Ok, I agree. Let's go with option 1 for now (and summit) and switch to option 2 in future when some of the problems are resolved. > but it does mean we have to provide some slave images, and/or an easy > way to create new slave images (which we actually already have here: > https://github.com/openshift/origin/blob/master/examples/jenkins/mast > er-slave/jenkins-slave-template.json), so certainly (2) feels like a > better long term solution. I have tried that template but if I understand it correctly: ?- it uses strategy "Docker" => privileged => won't work on Online and other secured environments? ?- requires your builder image to be CentOS based (which I am ok with - people could want something like fedora for newer packages, but they can always make the full image by themselves) Otherwise it can be useful. (Btw. if anyone is trying that template, you need to modify SLAVE_REPO_CONTEXTDIR?'s~examples/jenkins- master/slave~examples/jenkins/master-slave/slave~' - it used to live in another directory.) I guess, we should add this template to ADB, too. > > Providing maven/java and nodejs slave images out of the box seems > likely. Yes, when going with option 1, this is almost necessity. We will need at least a Java/Maven and NodeJS builders. And I was thinking they should be provided by default when you instantiate the new Jenkins template. What do you think, Ben? Or should Catapult be setting them up? The pipelines will be fully dependent on them. We should also coordinate on creating those basic images and publish them to Docker Hub, RH registry.? Maybe we could instantiate the template in OSE and set it up to push the images to Docker Hub? Thanks, Tomas > > ? > > > > > > > On 31 May 2016, at 15:19, Tomas Nozicka > > > wrote: > > > > > > Hi, > > > > > > how do we want to run our Jenkins "build slaves" in our > > > pipelines? > > > > > > Options: > > > 1) kubernetes plugin [1] > > > = Example = > > > node ('java8-maven') { > > > ? sh 'mvn package' > > > } > > > = EOF = > > > ?- tightly coupled to Jenkins instance (requires predefined pools > > > of > > > slaves which you can choose from) > > > ?- build environment is defined by admin not pipeline > > > (Jenkinsfile) > > > ?- not generic > > > > > > 2) kubernetes-workflow plugin [2] > > > = Example = > > > kubernetes.pod('buildpod').withImage('java8-maven').inside{ > > > ? sh 'mvn package' > > > } > > > =?EOF = > > > ?+ loosely coupled to Jenkins instance > > > ?+ pipeline (Jenkinsfile) specifies build environment > > > ?+ generic (pipelines can specify custom?images) > > > > > > I am for option 2 because of the loose coupling to Jenkins and > > > ability > > > to specify (custom) build images as part of Jenkinsfile. > > > > > > And AFAIK catapult project is partially about taking user > > > pipelines and > > > setting those up with OpenShift(+Jenkins), I just don't see it > > > working > > > with the tight coupling and option 1. > > > > > > If we decide to go with option 2 we should probably integrate > > > kubernetes-workflow plugin into jenkins-1-centos7:dev image. > > > > > > What do you think? > > > > > > Thanks, > > > Tomas > > > > > > [1] -?https://wiki.jenkins-ci.org/display/JENKINS/Kubernetes+Plug > > > in > > > [2] -?https://github.com/fabric8io/kubernetes-workflow > > > > > > _______________________________________________ > > > kontinuity-dev-public mailing list > > > kontinuity-dev-public at redhat.com > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > James > > ------- > > Red Hat > > > > Twitter: @jstrachan > > Email: jstracha at redhat.com > > Blog: https://medium.com/@jstrachan/ > > > > fabric8: http://fabric8.io/ > > open source microservices platform > > > > > > _______________________________________________ > > kontinuity-dev-public mailing list > > kontinuity-dev-public at redhat.com > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > From tnozicka at redhat.com Wed Jun 1 14:28:56 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Wed, 01 Jun 2016 16:28:56 +0200 Subject: [kontinuity-dev-public] "Build slaves" in Jenkins pipelines In-Reply-To: References: <1464704380.9457.63.camel@redhat.com> <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> <1464784671.3614.14.camel@redhat.com> Message-ID: <1464791336.3614.60.camel@redhat.com> Inline. On St, 2016-06-01 at 13:46 +0100, James Strachan wrote: > > On 1 Jun 2016, at 13:37, Tomas Nozicka wrote: > > > > Inline. > > > > On ?t, 2016-05-31 at 15:27 +0100, James Strachan wrote: > > > I?m a huge fan of 2); but currently that requires a pod for the > > > JNLP > > > slave and another pod for the docker image; which on premise or > > > on? > > James, why does it need 2 pods? Could you elaborate on that? > one for JNLP and one for the docker images listed in the > kubernetes.pod.withImage statement. Ok, I will rephrase: And this is probably because of me missing some context, but why do you need to create 2 pods when you can run JNLP image and builder image(s) in the same pod? Pod can contain multiple containers and with 'kubernetes.pod.withImage' you are only specifying one. > > > > I am thinking more in terms of 1 pod with two containers (JNLP and > > builder image) or one builder image with JNLP embedded. > > > > Is this something that can be fixed in future? > Any contributions most welcome! > https://github.com/fabric8io/kubernetes-workflow/issues/40 > > > > > > Thanks, > > Tomas > > > > > Dedicated doesn?t really matter. But on OpenShift Online that is > > > gonna be painful (3 pods just for builds!); so option 1) using > > > one or > > > two curated pools of slaves of custom pod/mages might be simpler > > > for > > > Online. e.g. one slave pool for maven/java and one for nodejs? > > > > > > > On 31 May 2016, at 15:19, Tomas Nozicka > > > > wrote: > > > > > > > > Hi, > > > > > > > > how do we want to run our Jenkins "build slaves" in our > > > > pipelines? > > > > > > > > Options: > > > > 1) kubernetes plugin [1] > > > > = Example = > > > > node ('java8-maven') { > > > > ? sh 'mvn package' > > > > } > > > > = EOF = > > > > ?- tightly coupled to Jenkins instance (requires predefined > > > > pools > > > > of > > > > slaves which you can choose from) > > > > ?- build environment is defined by admin not pipeline > > > > (Jenkinsfile) > > > > ?- not generic > > > > > > > > 2) kubernetes-workflow plugin [2] > > > > = Example = > > > > kubernetes.pod('buildpod').withImage('java8-maven').inside{ > > > > ? sh 'mvn package' > > > > } > > > > =?EOF = > > > > ?+ loosely coupled to Jenkins instance > > > > ?+ pipeline (Jenkinsfile) specifies build environment > > > > ?+ generic (pipelines can specify custom?images) > > > > > > > > I am for option 2 because of the loose coupling to Jenkins and > > > > ability > > > > to specify (custom) build images as part of Jenkinsfile. > > > > > > > > And AFAIK catapult project is partially about taking user > > > > pipelines > > > > and > > > > setting those up with OpenShift(+Jenkins), I just don't see it > > > > working > > > > with the tight coupling and option 1. > > > > > > > > If we decide to go with option 2 we should probably integrate > > > > kubernetes-workflow plugin into jenkins-1-centos7:dev image. > > > > > > > > What do you think? > > > > > > > > Thanks, > > > > Tomas > > > > > > > > [1] -?https://wiki.jenkins-ci.org/display/JENKINS/Kubernetes+Pl > > > > ugin > > > > [2] -?https://github.com/fabric8io/kubernetes-workflow > > > > > > > > _______________________________________________ > > > > kontinuity-dev-public mailing list > > > > kontinuity-dev-public at redhat.com > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > James > > > ------- > > > Red Hat > > > > > > Twitter: @jstrachan > > > Email: jstracha at redhat.com > > > Blog: https://medium.com/@jstrachan/ > > > > > > fabric8: http://fabric8.io/ > > > open source microservices platform > > > > > James > ------- > Red Hat > > Twitter: @jstrachan > Email: jstracha at redhat.com > Blog: https://medium.com/@jstrachan/ > > fabric8: http://fabric8.io/ > open source microservices platform > From jstracha at redhat.com Wed Jun 1 14:33:44 2016 From: jstracha at redhat.com (James Strachan) Date: Wed, 1 Jun 2016 15:33:44 +0100 Subject: [kontinuity-dev-public] "Build slaves" in Jenkins pipelines In-Reply-To: <1464791336.3614.60.camel@redhat.com> References: <1464704380.9457.63.camel@redhat.com> <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> <1464784671.3614.14.camel@redhat.com> <1464791336.3614.60.camel@redhat.com> Message-ID: <01DA73CE-A06A-43E0-A8AC-1895C233E3AA@redhat.com> > On 1 Jun 2016, at 15:28, Tomas Nozicka wrote: > > Inline. > > On St, 2016-06-01 at 13:46 +0100, James Strachan wrote: >>> On 1 Jun 2016, at 13:37, Tomas Nozicka wrote: >>> >>> Inline. >>> >>> On ?t, 2016-05-31 at 15:27 +0100, James Strachan wrote: >>>> I?m a huge fan of 2); but currently that requires a pod for the >>>> JNLP >>>> slave and another pod for the docker image; which on premise or >>>> on >>> James, why does it need 2 pods? Could you elaborate on that? >> one for JNLP and one for the docker images listed in the >> kubernetes.pod.withImage statement. > Ok, I will rephrase: > And this is probably because of me missing some context, but why do you > need to create 2 pods when you can run JNLP image and builder image(s) > in the same pod? Pod can contain multiple containers and with > 'kubernetes.pod.withImage' you are only specifying one. Technically I?m hoping that can be done - someone just needs to figure out how to do it ;) James ------- Red Hat Twitter: @jstrachan Email: jstracha at redhat.com Blog: https://medium.com/@jstrachan/ fabric8: http://fabric8.io/ open source microservices platform -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Wed Jun 1 23:42:46 2016 From: bparees at redhat.com (Ben Parees) Date: Wed, 1 Jun 2016 19:42:46 -0400 Subject: [kontinuity-dev-public] "Build slaves" in Jenkins pipelines In-Reply-To: <1464790084.3614.50.camel@redhat.com> References: <1464704380.9457.63.camel@redhat.com> <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> <1464790084.3614.50.camel@redhat.com> Message-ID: On Wed, Jun 1, 2016 at 10:08 AM, Tomas Nozicka wrote: > Inline. > > On ?t, 2016-05-31 at 22:58 -0400, Ben Parees wrote: > > > > > > On Tue, May 31, 2016 at 10:27 AM, James Strachan > > wrote: > > > I?m a huge fan of 2); but currently that requires a pod for the > > > JNLP slave and another pod for the docker image; which on premise > > > or on Dedicated doesn?t really matter. But on OpenShift Online that > > > is gonna be painful (3 pods just for builds!); so option 1) using > > > one or two curated pools of slaves of custom pod/mages might be > > > simpler for Online. e.g. one slave pool for maven/java and one for > > > nodejs? > > > > > yeah so i've been leaning towards (1) due to it needing fewer pods, > Ok, I agree. Let's go with option 1 for now (and summit) and switch to > option 2 in future when some of the problems are resolved. > > > but it does mean we have to provide some slave images, and/or an easy > > way to create new slave images (which we actually already have here: > > https://github.com/openshift/origin/blob/master/examples/jenkins/mast > > er-slave/jenkins-slave-template.json), so certainly (2) feels like a > > better long term solution. > I have tried that template but if I understand it correctly: > - it uses strategy "Docker" => privileged => won't work on Online and > other secured environments? > ?correct. ? > - requires your builder image to be CentOS based (which I am ok with - > people could want something like fedora for newer packages, but they > can always make the full image by themselves) > Otherwise it can be useful. > ?you can provide another dockerfile that's not centos based, the template is really just an example.? > > (Btw. if anyone is trying that template, you need to modify > SLAVE_REPO_CONTEXTDIR 's~examples/jenkins- > master/slave~examples/jenkins/master-slave/slave~' - it used to live in > another directory.) > ?thanks for pointing this out, fix coming: https://github.com/openshift/origin/pull/9123 ? > > I guess, we should add this template to ADB, too. > > > > > Providing maven/java and nodejs slave images out of the box seems > > likely. > Yes, when going with option 1, this is almost necessity. We will need > at least a Java/Maven and NodeJS builders. And I was thinking they > should be provided by default when you instantiate the new Jenkins > template. What do you think, Ben? > ?i meant make them available in dockerhub and registry.access.redhat.com (and possibly register imagestream definitions for them in the openhshift namespace by default). I don't know what "provide them when you instantiate the jenkins template" would mean.? > > Or should Catapult be setting them up? The pipelines will be fully > dependent on them. > > We should also coordinate on creating those basic images and publish > them to Docker Hub, RH registry. > > Maybe we could instantiate the template in OSE and set it up to push > the images to Docker Hub? > > Thanks, > Tomas > > > > > > > > > > > > > > > On 31 May 2016, at 15:19, Tomas Nozicka > > > > wrote: > > > > > > > > Hi, > > > > > > > > how do we want to run our Jenkins "build slaves" in our > > > > pipelines? > > > > > > > > Options: > > > > 1) kubernetes plugin [1] > > > > = Example = > > > > node ('java8-maven') { > > > > sh 'mvn package' > > > > } > > > > = EOF = > > > > - tightly coupled to Jenkins instance (requires predefined pools > > > > of > > > > slaves which you can choose from) > > > > - build environment is defined by admin not pipeline > > > > (Jenkinsfile) > > > > - not generic > > > > > > > > 2) kubernetes-workflow plugin [2] > > > > = Example = > > > > kubernetes.pod('buildpod').withImage('java8-maven').inside{ > > > > sh 'mvn package' > > > > } > > > > = EOF = > > > > + loosely coupled to Jenkins instance > > > > + pipeline (Jenkinsfile) specifies build environment > > > > + generic (pipelines can specify custom images) > > > > > > > > I am for option 2 because of the loose coupling to Jenkins and > > > > ability > > > > to specify (custom) build images as part of Jenkinsfile. > > > > > > > > And AFAIK catapult project is partially about taking user > > > > pipelines and > > > > setting those up with OpenShift(+Jenkins), I just don't see it > > > > working > > > > with the tight coupling and option 1. > > > > > > > > If we decide to go with option 2 we should probably integrate > > > > kubernetes-workflow plugin into jenkins-1-centos7:dev image. > > > > > > > > What do you think? > > > > > > > > Thanks, > > > > Tomas > > > > > > > > [1] - https://wiki.jenkins-ci.org/display/JENKINS/Kubernetes+Plug > > > > in > > > > [2] - https://github.com/fabric8io/kubernetes-workflow > > > > > > > > _______________________________________________ > > > > kontinuity-dev-public mailing list > > > > kontinuity-dev-public at redhat.com > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > > > James > > > ------- > > > Red Hat > > > > > > Twitter: @jstrachan > > > Email: jstracha at redhat.com > > > Blog: https://medium.com/@jstrachan/ > > > > > > fabric8: http://fabric8.io/ > > > open source microservices platform > > > > > > > > > _______________________________________________ > > > kontinuity-dev-public mailing list > > > kontinuity-dev-public at redhat.com > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > > > > > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Wed Jun 1 23:46:50 2016 From: bparees at redhat.com (Ben Parees) Date: Wed, 1 Jun 2016 19:46:50 -0400 Subject: [kontinuity-dev-public] "Build slaves" in Jenkins pipelines In-Reply-To: <01DA73CE-A06A-43E0-A8AC-1895C233E3AA@redhat.com> References: <1464704380.9457.63.camel@redhat.com> <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> <1464784671.3614.14.camel@redhat.com> <1464791336.3614.60.camel@redhat.com> <01DA73CE-A06A-43E0-A8AC-1895C233E3AA@redhat.com> Message-ID: On Wed, Jun 1, 2016 at 10:33 AM, James Strachan wrote: > > On 1 Jun 2016, at 15:28, Tomas Nozicka wrote: > > Inline. > > On St, 2016-06-01 at 13:46 +0100, James Strachan wrote: > > On 1 Jun 2016, at 13:37, Tomas Nozicka wrote: > > Inline. > > On ?t, 2016-05-31 at 15:27 +0100, James Strachan wrote: > > I?m a huge fan of 2); but currently that requires a pod for the > JNLP > slave and another pod for the docker image; which on premise or > on > > James, why does it need 2 pods? Could you elaborate on that? > > one for JNLP and one for the docker images listed in the > kubernetes.pod.withImage statement. > > Ok, I will rephrase: > And this is probably because of me missing some context, but why do you > need to create 2 pods when you can run JNLP image and builder image(s) > in the same pod? Pod can contain multiple containers and with > 'kubernetes.pod.withImage' you are only specifying one. > > > Technically I?m hoping that can be done - someone just needs to figure out > how to do it ;) > ?two contains still requires resource for each container so this may not help the online scenario as much as we'd like, though it is a more logical configuration. > James > ------- > Red Hat > > Twitter: @jstrachan > Email: jstracha at redhat.com > Blog: https://medium.com/@jstrachan/ > > fabric8: http://fabric8.io/ > open source microservices platform > > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Thu Jun 2 07:54:07 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Thu, 02 Jun 2016 09:54:07 +0200 Subject: [kontinuity-dev-public] "Build slaves" in Jenkins pipelines In-Reply-To: References: <1464704380.9457.63.camel@redhat.com> <98E99398-C5E5-4556-8C26-FB576253BF72@redhat.com> <1464790084.3614.50.camel@redhat.com> Message-ID: <1464854047.9457.93.camel@redhat.com> Inline. On St, 2016-06-01 at 19:42 -0400, Ben Parees wrote: > > > On Wed, Jun 1, 2016 at 10:08 AM, Tomas Nozicka > wrote: > > Inline. > > > > On ?t, 2016-05-31 at 22:58 -0400, Ben Parees wrote: > > > > > > > > > On Tue, May 31, 2016 at 10:27 AM, James Strachan > .com > > > > wrote: > > > > I?m a huge fan of 2); but currently that requires a pod for the > > > > JNLP slave and another pod for the docker image; which on > > premise > > > > or on Dedicated doesn?t really matter. But on OpenShift Online > > that > > > > is gonna be painful (3 pods just for builds!); so option 1) > > using > > > > one or two curated pools of slaves of custom pod/mages might be > > > > simpler for Online. e.g. one slave pool for maven/java and one > > for > > > > nodejs? > > > > > > > yeah so i've been leaning towards (1) due to it needing fewer > > pods, > > Ok, I agree. Let's go with option 1 for now (and summit) and switch > > to > > option 2 in future when some of the problems are resolved. > > > > > but it does mean we have to provide some slave images, and/or an > > easy > > > way to create new slave images (which we actually already have > > here: > > > https://github.com/openshift/origin/blob/master/examples/jenkins/ > > mast > > > er-slave/jenkins-slave-template.json), so certainly (2) feels > > like a > > > better long term solution. > > I have tried that template but if I understand it correctly: > > ?- it uses strategy "Docker" => privileged => won't work on Online > > and > > other secured environments? > correct. > ? > > ?- requires your builder image to be CentOS based (which I am ok > > with - > > people could want something like fedora for newer packages, but > > they > > can always make the full image by themselves) > > Otherwise it can be useful. > you can provide another dockerfile that's not centos based, the > template is really just an example. True. > > ? > > > > (Btw. if anyone is trying that template, you need to modify > > SLAVE_REPO_CONTEXTDIR?'s~examples/jenkins- > > master/slave~examples/jenkins/master-slave/slave~' - it used to > > live in > > another directory.) > thanks for pointing this out, fix coming: > https://github.com/openshift/origin/pull/9123 > ? > > > > I guess, we should add this template to ADB, too. > > > > > > > > Providing maven/java and nodejs slave images out of the box seems > > > likely. > > Yes, when going with option 1, this is almost necessity. We will > > need > > at least a Java/Maven and NodeJS builders. And I was thinking they > > should be provided by default when you instantiate the new Jenkins > > template. What do you think, Ben? > i meant make them available in dockerhub and > registry.access.redhat.com (and possibly register imagestream > definitions for them in the openhshift namespace by default).? I > don't know what "provide them when you instantiate the jenkins > template" would mean. Sorry, I should have been more clearer. What I would like us to do: ?- make those images available in Docker Hub and registry.access.redhat.com ?- (optionally) have them in default ImageStreams in openshift namespace. ?- in image jenkins-1-centos7:dev create configured "Kubernetes Pod Template" section for each of them > > ? > > > > Or should Catapult be setting them up? The pipelines will be fully > > dependent on them. > > > > We should also coordinate on creating those basic images and > > publish > > them to Docker Hub, RH registry.? > > > > Maybe we could instantiate the template in OSE and set it up to > > push > > the images to Docker Hub? > > > > Thanks, > > Tomas > > > > > > > > ? > > > > > > > > > > > > > On 31 May 2016, at 15:19, Tomas Nozicka > > > > > wrote: > > > > > > > > > > Hi, > > > > > > > > > > how do we want to run our Jenkins "build slaves" in our > > > > > pipelines? > > > > > > > > > > Options: > > > > > 1) kubernetes plugin [1] > > > > > = Example = > > > > > node ('java8-maven') { > > > > > ? sh 'mvn package' > > > > > } > > > > > = EOF = > > > > > ?- tightly coupled to Jenkins instance (requires predefined > > pools > > > > > of > > > > > slaves which you can choose from) > > > > > ?- build environment is defined by admin not pipeline > > > > > (Jenkinsfile) > > > > > ?- not generic > > > > > > > > > > 2) kubernetes-workflow plugin [2] > > > > > = Example = > > > > > kubernetes.pod('buildpod').withImage('java8-maven').inside{ > > > > > ? sh 'mvn package' > > > > > } > > > > > =?EOF = > > > > > ?+ loosely coupled to Jenkins instance > > > > > ?+ pipeline (Jenkinsfile) specifies build environment > > > > > ?+ generic (pipelines can specify custom?images) > > > > > > > > > > I am for option 2 because of the loose coupling to Jenkins > > and > > > > > ability > > > > > to specify (custom) build images as part of Jenkinsfile. > > > > > > > > > > And AFAIK catapult project is partially about taking user > > > > > pipelines and > > > > > setting those up with OpenShift(+Jenkins), I just don't see > > it > > > > > working > > > > > with the tight coupling and option 1. > > > > > > > > > > If we decide to go with option 2 we should probably integrate > > > > > kubernetes-workflow plugin into jenkins-1-centos7:dev image. > > > > > > > > > > What do you think? > > > > > > > > > > Thanks, > > > > > Tomas > > > > > > > > > > [1] -?https://wiki.jenkins-ci.org/display/JENKINS/Kubernetes+ > > Plug > > > > > in > > > > > [2] -?https://github.com/fabric8io/kubernetes-workflow > > > > > > > > > > _______________________________________________ > > > > > kontinuity-dev-public mailing list > > > > > kontinuity-dev-public at redhat.com > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > > > > > James > > > > ------- > > > > Red Hat > > > > > > > > Twitter: @jstrachan > > > > Email: jstracha at redhat.com > > > > Blog: https://medium.com/@jstrachan/ > > > > > > > > fabric8: http://fabric8.io/ > > > > open source microservices platform > > > > > > > > > > > > _______________________________________________ > > > > kontinuity-dev-public mailing list > > > > kontinuity-dev-public at redhat.com > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > > > > > > > > > > > From alr at redhat.com Thu Jun 2 22:26:19 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Thu, 2 Jun 2016 18:26:19 -0400 Subject: [kontinuity-dev-public] Visualizing Status for Pipeline Stages That Haven't Yet Started Message-ID: Let's shine a light on this one: https://github.com/fabric8io/openshift-jenkins-sync-plugin/issues/56 So the underlying issue is that the openshift-sync-plugin is not currently propagating information about future stages, preventing the OpenShift Console from building a view of the entire pipeline. A few questions: 1) Jimmi, would you mind running point for us to get an update on feasibility of a fix? 2) I'm available to work on this, but as noted in the comments can't find instructions to run this component or test it (aside from a couple of small unit tests); am I missing something that's outside the repo? So let's prioritize work here; folks are going to be excited to see visualized pipelines in OpenShift and 56 is a dependency. :) S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Thu Jun 2 23:36:29 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Thu, 2 Jun 2016 19:36:29 -0400 Subject: [kontinuity-dev-public] A Catapult Console Message-ID: Hi, all: Because our first release is going to be a local-only experience, we needed an entry point for users. That's now in the Catapult application in the form of an index.html: https://github.com/redhat-kontinuity/catapult/issues/108 It looks like: https://ibin.co/2jKJbAwPTLYO.png Simple form; you select one of a known collection of pipelineable repositories, and hit "Submit". Catapult will then take over with the: * GitHub login and OAuth process * Forking the selected repo into your GitHub namespace * Creating an OpenShift project for you * Redirecting you to the overview page for that project on the OpenShift console >From a pure Catapult perspective, the last piece to be done is being handled by Xavier: creating a pipeline build/app for the project in question (at the moment it just makes a new project). Feel free to use this in your development if relevant locally: http://127.0.0.1:8080/kontinuity-catapult S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From ffranz at redhat.com Fri Jun 3 14:42:16 2016 From: ffranz at redhat.com (Fabiano Franz) Date: Fri, 3 Jun 2016 11:42:16 -0300 Subject: [kontinuity-dev-public] Visualizing Status for Pipeline Stages That Haven't Yet Started In-Reply-To: References: Message-ID: +1 about 56, but in terms of priorities I'd say we *need* to address 70 first. The build sync is broken, nothing (status, stages, etc) is being propagated back to the build in OpenShift at the moment. Fabiano Franz On Thu, Jun 2, 2016 at 7:26 PM, Andrew Lee Rubinger wrote: > Let's shine a light on this one: > > https://github.com/fabric8io/openshift-jenkins-sync-plugin/issues/56 > > So the underlying issue is that the openshift-sync-plugin is not currently > propagating information about future stages, preventing the OpenShift > Console from building a view of the entire pipeline. > > A few questions: > > 1) Jimmi, would you mind running point for us to get an update on > feasibility of a fix? > 2) I'm available to work on this, but as noted in the comments can't find > instructions to run this component or test it (aside from a couple of small > unit tests); am I missing something that's outside the repo? > > So let's prioritize work here; folks are going to be excited to see > visualized pipelines in OpenShift and 56 is a dependency. :) > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jdyson at redhat.com Fri Jun 3 14:52:47 2016 From: jdyson at redhat.com (Jimmi Dyson) Date: Fri, 3 Jun 2016 15:52:47 +0100 Subject: [kontinuity-dev-public] Visualizing Status for Pipeline Stages That Haven't Yet Started In-Reply-To: References: Message-ID: Yeah I'm working on fixes for those right now. Not quite as simple as I would have hoped as we have to update the kubernetes client to handle PATCHes as opposed to the PUTs that it has always used successfully until now... On 3 June 2016 at 15:42, Fabiano Franz wrote: > +1 about 56, but in terms of priorities I'd say we *need* to address 70 > first. The build sync is broken, nothing (status, stages, etc) is being > propagated back to the build in OpenShift at the moment. > > Fabiano Franz > > > > > > > On Thu, Jun 2, 2016 at 7:26 PM, Andrew Lee Rubinger wrote: >> >> Let's shine a light on this one: >> >> https://github.com/fabric8io/openshift-jenkins-sync-plugin/issues/56 >> >> So the underlying issue is that the openshift-sync-plugin is not currently >> propagating information about future stages, preventing the OpenShift >> Console from building a view of the entire pipeline. >> >> A few questions: >> >> 1) Jimmi, would you mind running point for us to get an update on >> feasibility of a fix? >> 2) I'm available to work on this, but as noted in the comments can't find >> instructions to run this component or test it (aside from a couple of small >> unit tests); am I missing something that's outside the repo? >> >> So let's prioritize work here; folks are going to be excited to see >> visualized pipelines in OpenShift and 56 is a dependency. :) >> >> S, >> ALR >> >> -- >> Red Hat Developer Programs Architecture >> @ALRubinger >> >> _______________________________________________ >> kontinuity-dev-public mailing list >> kontinuity-dev-public at redhat.com >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> > From ffranz at redhat.com Fri Jun 3 15:03:00 2016 From: ffranz at redhat.com (Fabiano Franz) Date: Fri, 3 Jun 2016 12:03:00 -0300 Subject: [kontinuity-dev-public] Visualizing Status for Pipeline Stages That Haven't Yet Started In-Reply-To: References: Message-ID: Great, thanks Jimmi! On Fri, Jun 3, 2016 at 11:52 AM, Jimmi Dyson wrote: > Yeah I'm working on fixes for those right now. Not quite as simple as > I would have hoped as we have to update the kubernetes client to > handle PATCHes as opposed to the PUTs that it has always used > successfully until now... > > On 3 June 2016 at 15:42, Fabiano Franz wrote: > > +1 about 56, but in terms of priorities I'd say we *need* to address 70 > > first. The build sync is broken, nothing (status, stages, etc) is being > > propagated back to the build in OpenShift at the moment. > > > > Fabiano Franz > > > > > > > > > > > > > > On Thu, Jun 2, 2016 at 7:26 PM, Andrew Lee Rubinger > wrote: > >> > >> Let's shine a light on this one: > >> > >> https://github.com/fabric8io/openshift-jenkins-sync-plugin/issues/56 > >> > >> So the underlying issue is that the openshift-sync-plugin is not > currently > >> propagating information about future stages, preventing the OpenShift > >> Console from building a view of the entire pipeline. > >> > >> A few questions: > >> > >> 1) Jimmi, would you mind running point for us to get an update on > >> feasibility of a fix? > >> 2) I'm available to work on this, but as noted in the comments can't > find > >> instructions to run this component or test it (aside from a couple of > small > >> unit tests); am I missing something that's outside the repo? > >> > >> So let's prioritize work here; folks are going to be excited to see > >> visualized pipelines in OpenShift and 56 is a dependency. :) > >> > >> S, > >> ALR > >> > >> -- > >> Red Hat Developer Programs Architecture > >> @ALRubinger > >> > >> _______________________________________________ > >> kontinuity-dev-public mailing list > >> kontinuity-dev-public at redhat.com > >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > >> > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Fri Jun 3 15:38:37 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Fri, 3 Jun 2016 11:38:37 -0400 Subject: [kontinuity-dev-public] Visualizing Status for Pipeline Stages That Haven't Yet Started In-Reply-To: References: Message-ID: Thanks for the note on #70; added that one to my lurking list. :) I'll let ya hack away; let's check in top of the week to see where we're at -- and remember for any Java stuff I can help with, just get me instructions for build/test and I'm there. S, ALR On Fri, Jun 3, 2016 at 11:03 AM, Fabiano Franz wrote: > Great, thanks Jimmi! > > On Fri, Jun 3, 2016 at 11:52 AM, Jimmi Dyson wrote: > >> Yeah I'm working on fixes for those right now. Not quite as simple as >> I would have hoped as we have to update the kubernetes client to >> handle PATCHes as opposed to the PUTs that it has always used >> successfully until now... >> >> On 3 June 2016 at 15:42, Fabiano Franz wrote: >> > +1 about 56, but in terms of priorities I'd say we *need* to address 70 >> > first. The build sync is broken, nothing (status, stages, etc) is being >> > propagated back to the build in OpenShift at the moment. >> > >> > Fabiano Franz >> > >> > >> > >> > >> > >> > >> > On Thu, Jun 2, 2016 at 7:26 PM, Andrew Lee Rubinger >> wrote: >> >> >> >> Let's shine a light on this one: >> >> >> >> https://github.com/fabric8io/openshift-jenkins-sync-plugin/issues/56 >> >> >> >> So the underlying issue is that the openshift-sync-plugin is not >> currently >> >> propagating information about future stages, preventing the OpenShift >> >> Console from building a view of the entire pipeline. >> >> >> >> A few questions: >> >> >> >> 1) Jimmi, would you mind running point for us to get an update on >> >> feasibility of a fix? >> >> 2) I'm available to work on this, but as noted in the comments can't >> find >> >> instructions to run this component or test it (aside from a couple of >> small >> >> unit tests); am I missing something that's outside the repo? >> >> >> >> So let's prioritize work here; folks are going to be excited to see >> >> visualized pipelines in OpenShift and 56 is a dependency. :) >> >> >> >> S, >> >> ALR >> >> >> >> -- >> >> Red Hat Developer Programs Architecture >> >> @ALRubinger >> >> >> >> _______________________________________________ >> >> kontinuity-dev-public mailing list >> >> kontinuity-dev-public at redhat.com >> >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> >> >> > >> > > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From jdyson at redhat.com Fri Jun 3 15:49:28 2016 From: jdyson at redhat.com (Jimmi Dyson) Date: Fri, 3 Jun 2016 16:49:28 +0100 Subject: [kontinuity-dev-public] Visualizing Status for Pipeline Stages That Haven't Yet Started In-Reply-To: References: Message-ID: Cool! Instructions for local dev: 1. `oc login` to user with enough privileges to update build configs/builds (I use cluster-admin for ease). 2. Run `mvn hpi:run` to start dev Jenkins with required plugins 3. Jenkins will be running at http://localhost:8080/jenkins 4. Hack whatever you need - totally welcome to look at 56 while I work on 70. Testing is all manual right now - yes I know this is bad but it;s just the way it is right now. Anyone fancy adding some integration tests? :) On 3 June 2016 at 16:38, Andrew Lee Rubinger wrote: > Thanks for the note on #70; added that one to my lurking list. :) > > I'll let ya hack away; let's check in top of the week to see where we're at > -- and remember for any Java stuff I can help with, just get me instructions > for build/test and I'm there. > > S, > ALR > > On Fri, Jun 3, 2016 at 11:03 AM, Fabiano Franz wrote: >> >> Great, thanks Jimmi! >> >> On Fri, Jun 3, 2016 at 11:52 AM, Jimmi Dyson wrote: >>> >>> Yeah I'm working on fixes for those right now. Not quite as simple as >>> I would have hoped as we have to update the kubernetes client to >>> handle PATCHes as opposed to the PUTs that it has always used >>> successfully until now... >>> >>> On 3 June 2016 at 15:42, Fabiano Franz wrote: >>> > +1 about 56, but in terms of priorities I'd say we *need* to address 70 >>> > first. The build sync is broken, nothing (status, stages, etc) is being >>> > propagated back to the build in OpenShift at the moment. >>> > >>> > Fabiano Franz >>> > >>> > >>> > >>> > >>> > >>> > >>> > On Thu, Jun 2, 2016 at 7:26 PM, Andrew Lee Rubinger >>> > wrote: >>> >> >>> >> Let's shine a light on this one: >>> >> >>> >> https://github.com/fabric8io/openshift-jenkins-sync-plugin/issues/56 >>> >> >>> >> So the underlying issue is that the openshift-sync-plugin is not >>> >> currently >>> >> propagating information about future stages, preventing the OpenShift >>> >> Console from building a view of the entire pipeline. >>> >> >>> >> A few questions: >>> >> >>> >> 1) Jimmi, would you mind running point for us to get an update on >>> >> feasibility of a fix? >>> >> 2) I'm available to work on this, but as noted in the comments can't >>> >> find >>> >> instructions to run this component or test it (aside from a couple of >>> >> small >>> >> unit tests); am I missing something that's outside the repo? >>> >> >>> >> So let's prioritize work here; folks are going to be excited to see >>> >> visualized pipelines in OpenShift and 56 is a dependency. :) >>> >> >>> >> S, >>> >> ALR >>> >> >>> >> -- >>> >> Red Hat Developer Programs Architecture >>> >> @ALRubinger >>> >> >>> >> _______________________________________________ >>> >> kontinuity-dev-public mailing list >>> >> kontinuity-dev-public at redhat.com >>> >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >>> >> >>> > >> >> > > > > -- > Red Hat Developer Programs Architecture > @ALRubinger From alr at redhat.com Fri Jun 3 22:33:31 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Fri, 3 Jun 2016 18:33:31 -0400 Subject: [kontinuity-dev-public] Visualizing Status for Pipeline Stages That Haven't Yet Started In-Reply-To: References: Message-ID: Thanks for that! I'd like to dig in this weekend. On first run of "mvn clean install": [ERROR] Failed to execute goal on project openshift-jenkins-sync-plugin: Could not resolve dependencies for project io.fabric8:openshift-jenkins-sync-plugin:hpi:0.1-SNAPSHOT: Could not find artifact io.fabric8:kubernetes-model:jar:1.0.44-SNAPSHOT in repo.jenkins-ci.org (http://repo.jenkins-ci.org/public/) -> [Help 1] Is there some other remote repo which is a dependency or do I need to build kubernetes-model on my own and install into my local Maven repo? S, ALR On Fri, Jun 3, 2016 at 11:49 AM, Jimmi Dyson wrote: > Cool! Instructions for local dev: > > 1. `oc login` to user with enough privileges to update build > configs/builds (I use cluster-admin for ease). > 2. Run `mvn hpi:run` to start dev Jenkins with required plugins > 3. Jenkins will be running at http://localhost:8080/jenkins > 4. Hack whatever you need - totally welcome to look at 56 while I work on > 70. > > Testing is all manual right now - yes I know this is bad but it;s just > the way it is right now. Anyone fancy adding some integration tests? > :) > > On 3 June 2016 at 16:38, Andrew Lee Rubinger wrote: > > Thanks for the note on #70; added that one to my lurking list. :) > > > > I'll let ya hack away; let's check in top of the week to see where we're > at > > -- and remember for any Java stuff I can help with, just get me > instructions > > for build/test and I'm there. > > > > S, > > ALR > > > > On Fri, Jun 3, 2016 at 11:03 AM, Fabiano Franz > wrote: > >> > >> Great, thanks Jimmi! > >> > >> On Fri, Jun 3, 2016 at 11:52 AM, Jimmi Dyson wrote: > >>> > >>> Yeah I'm working on fixes for those right now. Not quite as simple as > >>> I would have hoped as we have to update the kubernetes client to > >>> handle PATCHes as opposed to the PUTs that it has always used > >>> successfully until now... > >>> > >>> On 3 June 2016 at 15:42, Fabiano Franz wrote: > >>> > +1 about 56, but in terms of priorities I'd say we *need* to address > 70 > >>> > first. The build sync is broken, nothing (status, stages, etc) is > being > >>> > propagated back to the build in OpenShift at the moment. > >>> > > >>> > Fabiano Franz > >>> > > >>> > > >>> > > >>> > > >>> > > >>> > > >>> > On Thu, Jun 2, 2016 at 7:26 PM, Andrew Lee Rubinger > >>> > wrote: > >>> >> > >>> >> Let's shine a light on this one: > >>> >> > >>> >> > https://github.com/fabric8io/openshift-jenkins-sync-plugin/issues/56 > >>> >> > >>> >> So the underlying issue is that the openshift-sync-plugin is not > >>> >> currently > >>> >> propagating information about future stages, preventing the > OpenShift > >>> >> Console from building a view of the entire pipeline. > >>> >> > >>> >> A few questions: > >>> >> > >>> >> 1) Jimmi, would you mind running point for us to get an update on > >>> >> feasibility of a fix? > >>> >> 2) I'm available to work on this, but as noted in the comments can't > >>> >> find > >>> >> instructions to run this component or test it (aside from a couple > of > >>> >> small > >>> >> unit tests); am I missing something that's outside the repo? > >>> >> > >>> >> So let's prioritize work here; folks are going to be excited to see > >>> >> visualized pipelines in OpenShift and 56 is a dependency. :) > >>> >> > >>> >> S, > >>> >> ALR > >>> >> > >>> >> -- > >>> >> Red Hat Developer Programs Architecture > >>> >> @ALRubinger > >>> >> > >>> >> _______________________________________________ > >>> >> kontinuity-dev-public mailing list > >>> >> kontinuity-dev-public at redhat.com > >>> >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > >>> >> > >>> > > >> > >> > > > > > > > > -- > > Red Hat Developer Programs Architecture > > @ALRubinger > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Sat Jun 4 00:03:35 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Fri, 3 Jun 2016 20:03:35 -0400 Subject: [kontinuity-dev-public] Visualizing Status for Pipeline Stages That Haven't Yet Started In-Reply-To: References: Message-ID: And to add some more context: If I manually bump the version from 1.0.44-SNAPSHOT to 1.0.44, we get a compilation error: [ERROR] /home/alr/business/jboss/git/fabric8io/openshift-jenkins-sync-plugin/src/main/java/io/fabric8/jenkins/openshiftsync/BuildConfigToJobMapper.java:[29,38] cannot find symbol [ERROR] symbol: class JenkinsPipelineBuildStrategy [ERROR] location: package io.fabric8.openshift.api.model I don't know where JenkinsPipelineBuildStrategy is in kubernetes-model 1.0.44 or in upstream/master; appears to be gone? Totally open to working on this, but need a buildable thing. :) Let's try to get me ramped up next week so I can help out! S, ALR On Fri, Jun 3, 2016 at 6:33 PM, Andrew Lee Rubinger wrote: > Thanks for that! I'd like to dig in this weekend. > > On first run of "mvn clean install": > > [ERROR] Failed to execute goal on project openshift-jenkins-sync-plugin: > Could not resolve dependencies for project > io.fabric8:openshift-jenkins-sync-plugin:hpi:0.1-SNAPSHOT: Could not find > artifact io.fabric8:kubernetes-model:jar:1.0.44-SNAPSHOT in > repo.jenkins-ci.org (http://repo.jenkins-ci.org/public/) -> [Help 1] > > Is there some other remote repo which is a dependency or do I need to > build kubernetes-model on my own and install into my local Maven repo? > > S, > ALR > > > On Fri, Jun 3, 2016 at 11:49 AM, Jimmi Dyson wrote: > >> Cool! Instructions for local dev: >> >> 1. `oc login` to user with enough privileges to update build >> configs/builds (I use cluster-admin for ease). >> 2. Run `mvn hpi:run` to start dev Jenkins with required plugins >> 3. Jenkins will be running at http://localhost:8080/jenkins >> 4. Hack whatever you need - totally welcome to look at 56 while I work on >> 70. >> >> Testing is all manual right now - yes I know this is bad but it;s just >> the way it is right now. Anyone fancy adding some integration tests? >> :) >> >> On 3 June 2016 at 16:38, Andrew Lee Rubinger wrote: >> > Thanks for the note on #70; added that one to my lurking list. :) >> > >> > I'll let ya hack away; let's check in top of the week to see where >> we're at >> > -- and remember for any Java stuff I can help with, just get me >> instructions >> > for build/test and I'm there. >> > >> > S, >> > ALR >> > >> > On Fri, Jun 3, 2016 at 11:03 AM, Fabiano Franz >> wrote: >> >> >> >> Great, thanks Jimmi! >> >> >> >> On Fri, Jun 3, 2016 at 11:52 AM, Jimmi Dyson >> wrote: >> >>> >> >>> Yeah I'm working on fixes for those right now. Not quite as simple as >> >>> I would have hoped as we have to update the kubernetes client to >> >>> handle PATCHes as opposed to the PUTs that it has always used >> >>> successfully until now... >> >>> >> >>> On 3 June 2016 at 15:42, Fabiano Franz wrote: >> >>> > +1 about 56, but in terms of priorities I'd say we *need* to >> address 70 >> >>> > first. The build sync is broken, nothing (status, stages, etc) is >> being >> >>> > propagated back to the build in OpenShift at the moment. >> >>> > >> >>> > Fabiano Franz >> >>> > >> >>> > >> >>> > >> >>> > >> >>> > >> >>> > >> >>> > On Thu, Jun 2, 2016 at 7:26 PM, Andrew Lee Rubinger > > >> >>> > wrote: >> >>> >> >> >>> >> Let's shine a light on this one: >> >>> >> >> >>> >> >> https://github.com/fabric8io/openshift-jenkins-sync-plugin/issues/56 >> >>> >> >> >>> >> So the underlying issue is that the openshift-sync-plugin is not >> >>> >> currently >> >>> >> propagating information about future stages, preventing the >> OpenShift >> >>> >> Console from building a view of the entire pipeline. >> >>> >> >> >>> >> A few questions: >> >>> >> >> >>> >> 1) Jimmi, would you mind running point for us to get an update on >> >>> >> feasibility of a fix? >> >>> >> 2) I'm available to work on this, but as noted in the comments >> can't >> >>> >> find >> >>> >> instructions to run this component or test it (aside from a couple >> of >> >>> >> small >> >>> >> unit tests); am I missing something that's outside the repo? >> >>> >> >> >>> >> So let's prioritize work here; folks are going to be excited to see >> >>> >> visualized pipelines in OpenShift and 56 is a dependency. :) >> >>> >> >> >>> >> S, >> >>> >> ALR >> >>> >> >> >>> >> -- >> >>> >> Red Hat Developer Programs Architecture >> >>> >> @ALRubinger >> >>> >> >> >>> >> _______________________________________________ >> >>> >> kontinuity-dev-public mailing list >> >>> >> kontinuity-dev-public at redhat.com >> >>> >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> >>> >> >> >>> > >> >> >> >> >> > >> > >> > >> > -- >> > Red Hat Developer Programs Architecture >> > @ALRubinger >> > > > > -- > Red Hat Developer Programs Architecture > @ALRubinger > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Mon Jun 6 13:08:50 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Mon, 06 Jun 2016 15:08:50 +0200 Subject: [kontinuity-dev-public] Jenkins slaves can not be created from ImageStreams with jenkins-k8s-plugin Message-ID: <1465218530.3805.30.camel@redhat.com> I have been trying to create some Jenkins slaves from Ben's slave template [1] to work with jenkins-k8s-plugin, but when I set up the ImageStreams (+BuildConfigs) to convert builder images into Jenkins slaves I can't get Kubernetes Pod Template to reference them. It takes only an image as an argument, not the ImageStream. (Which is ok from the point of the plugin since k8s doesn't have ImageStreams.) So I have tried to reference the ImageStream in the internal registry as an image (docker-registry.default:5000/openshift/maven-jenkins-slave) but that fails because the pod created by the jenkins-k8s-plugin is not authenticated to access the registry. The only way I got this working is to set up the BuildConfig, login to the internal registry manually, pull the image, re-tag it, push it to Docker Hub and set up Kubernetes Pod Template to reference the image on Docker Hub. This leaves us with just pre-configred/pre-created slaves and no way to easily create custom ones. Is there a way to let the jenkins-k8s-plugin to authenticate to the (OpenShift's) secure registry? Or another solution? Thanks, Tomas [1] -?https://github.com/openshift/origin/blob/master/examples/jenkins/ master-slave/jenkins-slave-template.json From bparees at redhat.com Mon Jun 6 14:20:29 2016 From: bparees at redhat.com (Ben Parees) Date: Mon, 6 Jun 2016 10:20:29 -0400 Subject: [kontinuity-dev-public] Jenkins slaves can not be created from ImageStreams with jenkins-k8s-plugin In-Reply-To: <1465218530.3805.30.camel@redhat.com> References: <1465218530.3805.30.camel@redhat.com> Message-ID: On Mon, Jun 6, 2016 at 9:08 AM, Tomas Nozicka wrote: > I have been trying to create some Jenkins slaves from Ben's slave > template [1] to work with jenkins-k8s-plugin, but when I set up the > ImageStreams (+BuildConfigs) to convert builder images into Jenkins > slaves I can't get Kubernetes Pod Template to reference them. It takes > only an image as an argument, not the ImageStream. (Which is ok from > the point of the plugin since k8s doesn't have ImageStreams.) So I have > tried to reference the ImageStream in the internal registry as an image > (docker-registry.default:5000/openshift/maven-jenkins-slave) but that > fails because the pod created by the jenkins-k8s-plugin is not > authenticated to access the registry. > > The only way I got this working is to set up the BuildConfig, login to > the internal registry manually, pull the image, re-tag it, push it to > Docker Hub and set up Kubernetes Pod Template to reference the image on > Docker Hub. > > This leaves us with just pre-configred/pre-created slaves and no way to > easily create custom ones. > > Is there a way to let the jenkins-k8s-plugin to authenticate to the > (OpenShift's) secure registry? Or another solution? > ?the service account being used by the slave pod needs to have permissions to pull the image in question, so depending on: 1) what service account the slave is running as (should be the default svc account for your namespace i think) 2) what namespace the slave image was pushed to you'll need to grant the right image-puller role for the right namespace to the right service account. ? > > > Thanks, > Tomas > > [1] - https://github.com/openshift/origin/blob/master/examples/jenkins/ > master-slave/jenkins-slave-template.json > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Mon Jun 6 16:48:44 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Mon, 06 Jun 2016 18:48:44 +0200 Subject: [kontinuity-dev-public] Jenkins slaves can not be created from ImageStreams with jenkins-k8s-plugin In-Reply-To: References: <1465218530.3805.30.camel@redhat.com> Message-ID: <1465231724.3805.47.camel@redhat.com> On Po, 2016-06-06 at 10:20 -0400, Ben Parees wrote: > > > On Mon, Jun 6, 2016 at 9:08 AM, Tomas Nozicka > wrote: > > I have been trying to create some Jenkins slaves from Ben's slave > > template [1] to work with jenkins-k8s-plugin, but when I set up the > > ImageStreams (+BuildConfigs) to convert builder images into Jenkins > > slaves I can't get Kubernetes Pod Template to reference them. It > > takes > > only an image as an argument, not the ImageStream. (Which is ok > > from > > the point of the plugin since k8s doesn't have ImageStreams.) So I > > have > > tried to reference the ImageStream in the internal registry as an > > image > > (docker-registry.default:5000/openshift/maven-jenkins-slave) but > > that > > fails because the pod created by the jenkins-k8s-plugin is not > > authenticated to access the registry. > > > > The only way I got this working is to set up the BuildConfig, login > > to > > the internal registry manually, pull the image, re-tag it, push it > > to > > Docker Hub and set up Kubernetes Pod Template to reference the > > image on > > Docker Hub. > > > > This leaves us with just pre-configred/pre-created slaves and no > > way to > > easily create custom ones. > > > > Is there a way to let the jenkins-k8s-plugin to authenticate to the > > (OpenShift's) secure registry? Or another solution? > the service account being used by the slave pod needs to have > permissions to pull the image in question, so depending on: > 1) what service account the slave is running as (should be the > default svc account for your namespace i think) > 2) what namespace the slave image was pushed to > > you'll need to grant the right image-puller role for the right > namespace to the right service account. Thanks Ben, but after trying to set up those permission I just noticed that this was a DNS failure. The error message states that this is because of credentials missing but in fact at the and I missed "no such host". The reason is the cluster DNS does not work on nodes (kubelet). This makes it unpleasant for setup since the svc IP of docker- registry.default differs every time. But I guess our pre-build images e.g. maven and node are going to be in docker hub so this will not affect them. It will affect only setting up custom builder images. > Failed to pull image "docker-registry.default/openshift/maven-jenkins- slave": image pull failed for docker-registry.default/openshift/maven- jenkins-slave, this may be because there are no credentials on this request. details: (unable to ping registry endpoint https://docker-regi stry.default/v0/ v2 ping attempt failed with error: Get https://docker- registry.default/v2/: dial tcp: lookup docker-registry.default: no such host v1 ping attempt failed with error: Get https://docker- registry.default/v1/_ping: dial tcp: lookup docker-registry.default: no such host) > > ? > > > > > > Thanks, > > Tomas > > > > [1] -?https://github.com/openshift/origin/blob/master/examples/jenk > > ins/ > > master-slave/jenkins-slave-template.json > > > > From bparees at redhat.com Mon Jun 6 16:53:20 2016 From: bparees at redhat.com (Ben Parees) Date: Mon, 6 Jun 2016 12:53:20 -0400 Subject: [kontinuity-dev-public] Jenkins slaves can not be created from ImageStreams with jenkins-k8s-plugin In-Reply-To: <1465231724.3805.47.camel@redhat.com> References: <1465218530.3805.30.camel@redhat.com> <1465231724.3805.47.camel@redhat.com> Message-ID: On Mon, Jun 6, 2016 at 12:48 PM, Tomas Nozicka wrote: > On Po, 2016-06-06 at 10:20 -0400, Ben Parees wrote: > > > > > > On Mon, Jun 6, 2016 at 9:08 AM, Tomas Nozicka > > wrote: > > > I have been trying to create some Jenkins slaves from Ben's slave > > > template [1] to work with jenkins-k8s-plugin, but when I set up the > > > ImageStreams (+BuildConfigs) to convert builder images into Jenkins > > > slaves I can't get Kubernetes Pod Template to reference them. It > > > takes > > > only an image as an argument, not the ImageStream. (Which is ok > > > from > > > the point of the plugin since k8s doesn't have ImageStreams.) So I > > > have > > > tried to reference the ImageStream in the internal registry as an > > > image > > > (docker-registry.default:5000/openshift/maven-jenkins-slave) but > > > that > > > fails because the pod created by the jenkins-k8s-plugin is not > > > authenticated to access the registry. > > > > > > The only way I got this working is to set up the BuildConfig, login > > > to > > > the internal registry manually, pull the image, re-tag it, push it > > > to > > > Docker Hub and set up Kubernetes Pod Template to reference the > > > image on > > > Docker Hub. > > > > > > This leaves us with just pre-configred/pre-created slaves and no > > > way to > > > easily create custom ones. > > > > > > Is there a way to let the jenkins-k8s-plugin to authenticate to the > > > (OpenShift's) secure registry? Or another solution? > > the service account being used by the slave pod needs to have > > permissions to pull the image in question, so depending on: > > 1) what service account the slave is running as (should be the > > default svc account for your namespace i think) > > 2) what namespace the slave image was pushed to > > > > you'll need to grant the right image-puller role for the right > > namespace to the right service account. > Thanks Ben, but after trying to set up those permission I just noticed > that this was a DNS failure. The error message states that this is > because of credentials missing but in fact at the and I missed "no such > host". > The reason is the cluster DNS does not work on nodes (kubelet). This > makes it unpleasant for setup since the svc IP of docker- > registry.default differs every time. > ?you can add the cluster DNS to the node's resolv.conf. > But I guess our pre-build images e.g. maven and node are going to be in > docker hub so this will not affect them. It will affect only setting up > custom builder images. > > > > > Failed to pull image "docker-registry.default/openshift/maven-jenkins- > slave": image pull failed for docker-registry.default/openshift/maven- > jenkins-slave, this may be because there are no credentials on this > request. details: (unable to ping registry endpoint https://docker-regi > stry.default/v0/ v2 ping attempt failed with error: Get https://docker- > registry.default/v2/: dial tcp: lookup docker-registry.default: no such > host v1 ping attempt failed with error: Get https://docker- > registry.default/v1/_ping: dial tcp: lookup docker-registry.default: no > such host) > > > > > > > > > > > > > > Thanks, > > > Tomas > > > > > > [1] - https://github.com/openshift/origin/blob/master/examples/jenk > > > ins/ > > > master-slave/jenkins-slave-template.json > > > > > > > > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Mon Jun 6 17:13:29 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Mon, 06 Jun 2016 19:13:29 +0200 Subject: [kontinuity-dev-public] Jenkins slaves can not be created from ImageStreams with jenkins-k8s-plugin In-Reply-To: References: <1465218530.3805.30.camel@redhat.com> <1465231724.3805.47.camel@redhat.com> Message-ID: <1465233209.3805.63.camel@redhat.com> On Po, 2016-06-06 at 12:53 -0400, Ben Parees wrote: > > > On Mon, Jun 6, 2016 at 12:48 PM, Tomas Nozicka > wrote: > > Thanks Ben, but after trying to set up those permission I just > > noticed > > that this was a DNS failure. The error message states that this is > > because of credentials missing but in fact at the and I missed "no > > such > > host". > > The reason is the cluster DNS does not work on nodes (kubelet). > > This > > makes it unpleasant for setup since the svc IP of docker- > > registry.default differs every time. > you can add the cluster DNS to the node's resolv.conf. I could. But resolv.conf is actually generated on some platform by NetworkManager and I don't want my instructions to start by several lines of setting it up for DNS. If it's not the default behaviour in OpenShift, then ok I will use IP.? It would be easier if the registry and even router had fixed IP by default every time their services get created in OpenShift? Just like kubernetes.default always has?172.30.0.1 > > But I guess our pre-build images e.g. maven and node are going to > > be in > > docker hub so this will not affect them. It will affect only > > setting up > > custom builder images. Regards, Tomas From alr at redhat.com Tue Jun 7 07:53:15 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Tue, 7 Jun 2016 03:53:15 -0400 Subject: [kontinuity-dev-public] Running Catapult Inside OpenShift / ADB from a Docker Image Message-ID: Dharmit and Andrew Block now got our CI environment publishing the Catapult application image to a temporary location on Docker Hub! I've started on work to get user instructions so that our community can run this on their own; this is the *distributable* piece we've been wanting for Summit start to come together. :) Looks a little like: * Install ADB * Use Vagrant file: ** curl https://raw.githubusercontent.com/ALRubinger/adb-atomic-developer-bundle/origin-latest/components/centos/centos-openshift-setup/Vagrantfile * vagrant up * vagrant ssh * oc login (admin/admin) * oc new-project catapult * oc new-app sabre1041/catapult:latest * oc status * Login to Origin console (https://10.1.2.2:8443/console) as admin/admin * Create route for Catapult * Check it out in the browser ** eg: http://catapult-catapult.centos7-adb.10.1.2.2.xip.io/kontinuity-catapult/ Currently throwing errors due to missing envvars, but that's fine (and expected). Also I'm working on making a template file so we can do away with the manual "oc" commands: https://github.com/redhat-kontinuity/catapult/issues/114 The live doc (viewable to RHT employees) I'm using to track this is: https://docs.google.com/a/redhat.com/document/d/1FtBt8XwZ237Mko6zEpkTOuFFhHfXM9THerjh702XA0c/edit?usp=sharing S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: screenshot1.png Type: image/png Size: 30332 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: screenshot2.png Type: image/png Size: 21475 bytes Desc: not available URL: From alr at redhat.com Tue Jun 7 09:12:47 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Tue, 7 Jun 2016 05:12:47 -0400 Subject: [kontinuity-dev-public] What's the IP Address of OpenShift Relative to a Running Pod? Message-ID: Quick question in $subject. Was expecting some OPENSHIFT_ADDRESS or something environment variable passed into the Docker container when it started but I don't see it listed. S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmorales at redhat.com Tue Jun 7 11:02:52 2016 From: jmorales at redhat.com (Jorge Morales Pou) Date: Tue, 7 Jun 2016 13:02:52 +0200 Subject: [kontinuity-dev-public] What's the IP Address of OpenShift Relative to a Running Pod? In-Reply-To: References: Message-ID: http://stackoverflow.com/questions/31557932/how-to-get-the-namespace-from-inside-a-pod-in-openshift Jorge Morales OpenShift by Red Hat EMEA Field Product Manager Product Marketing Manager On Tue, Jun 7, 2016 at 11:12 AM, Andrew Lee Rubinger wrote: > Quick question in $subject. Was expecting some OPENSHIFT_ADDRESS or > something environment variable passed into the Docker container when it > started but I don't see it listed. > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Tue Jun 7 13:04:11 2016 From: bparees at redhat.com (Ben Parees) Date: Tue, 7 Jun 2016 09:04:11 -0400 Subject: [kontinuity-dev-public] What's the IP Address of OpenShift Relative to a Running Pod? In-Reply-To: References: Message-ID: Not sure when it was added, but you can also look at: /var/run/secrets/kubernetes.io/serviceaccount/namespace to find the namespace. On Tue, Jun 7, 2016 at 7:02 AM, Jorge Morales Pou wrote: > > http://stackoverflow.com/questions/31557932/how-to-get-the-namespace-from-inside-a-pod-in-openshift > > Jorge Morales > OpenShift by Red Hat > EMEA Field Product Manager > Product Marketing Manager > > On Tue, Jun 7, 2016 at 11:12 AM, Andrew Lee Rubinger > wrote: > >> Quick question in $subject. Was expecting some OPENSHIFT_ADDRESS or >> something environment variable passed into the Docker container when it >> started but I don't see it listed. >> >> S, >> ALR >> >> -- >> Red Hat Developer Programs Architecture >> @ALRubinger >> >> _______________________________________________ >> kontinuity-dev-public mailing list >> kontinuity-dev-public at redhat.com >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> >> > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmorales at redhat.com Tue Jun 7 13:31:30 2016 From: jmorales at redhat.com (Jorge Morales Pou) Date: Tue, 7 Jun 2016 15:31:30 +0200 Subject: [kontinuity-dev-public] What's the IP Address of OpenShift Relative to a Running Pod? In-Reply-To: References: Message-ID: You can: TOKEN="$(cat /var/run/secrets/kubernetes.io/serviceaccount/token)" NAMESPACE="$(cat /var/run/secrets/kubernetes.io/serviceaccount/namespace)" POD_NAME="$(echo $HOSTNAME)" $ curl --cacert /var/run/secrets/kubernetes.io/serviceaccount/ca.crt \ " https://openshift.default.svc.cluster.local/api/v1/namespaces/${NAMESPACE}/pods/${POD_NAME}" \ -H "Authorization: Bearer $TOKEN" And then you get the json where you can query for ".status.podIP" Not really fancy. Jorge Morales OpenShift by Red Hat EMEA Field Product Manager Product Marketing Manager On Tue, Jun 7, 2016 at 3:04 PM, Ben Parees wrote: > Not sure when it was added, but you can also look at: > /var/run/secrets/kubernetes.io/serviceaccount/namespace > > to find the namespace. > > > On Tue, Jun 7, 2016 at 7:02 AM, Jorge Morales Pou > wrote: > >> >> http://stackoverflow.com/questions/31557932/how-to-get-the-namespace-from-inside-a-pod-in-openshift >> >> Jorge Morales >> OpenShift by Red Hat >> EMEA Field Product Manager >> Product Marketing Manager >> >> On Tue, Jun 7, 2016 at 11:12 AM, Andrew Lee Rubinger >> wrote: >> >>> Quick question in $subject. Was expecting some OPENSHIFT_ADDRESS or >>> something environment variable passed into the Docker container when it >>> started but I don't see it listed. >>> >>> S, >>> ALR >>> >>> -- >>> Red Hat Developer Programs Architecture >>> @ALRubinger >>> >>> _______________________________________________ >>> kontinuity-dev-public mailing list >>> kontinuity-dev-public at redhat.com >>> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >>> >>> >> >> _______________________________________________ >> kontinuity-dev-public mailing list >> kontinuity-dev-public at redhat.com >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> >> > > > -- > Ben Parees | OpenShift > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Tue Jun 7 15:43:39 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Tue, 7 Jun 2016 11:43:39 -0400 Subject: [kontinuity-dev-public] What's the IP Address of OpenShift Relative to a Running Pod? In-Reply-To: References: Message-ID: So basically "the local hostname 'openshift.default.svc.cluster.local'". :) Works; thanks guys! S, ALR On Tue, Jun 7, 2016 at 9:31 AM, Jorge Morales Pou wrote: > You can: > > TOKEN="$(cat /var/run/secrets/kubernetes.io/serviceaccount/token)" > NAMESPACE="$(cat /var/run/secrets/kubernetes.io/serviceaccount/namespace)" > POD_NAME="$(echo $HOSTNAME)" > > > $ curl --cacert /var/run/secrets/kubernetes.io/serviceaccount/ca.crt \ > " > https://openshift.default.svc.cluster.local/api/v1/namespaces/${NAMESPACE}/pods/${POD_NAME}" > \ > -H "Authorization: Bearer $TOKEN" > > And then you get the json where you can query for ".status.podIP" > > Not really fancy. > > Jorge Morales > OpenShift by Red Hat > EMEA Field Product Manager > Product Marketing Manager > > On Tue, Jun 7, 2016 at 3:04 PM, Ben Parees wrote: > >> Not sure when it was added, but you can also look at: >> /var/run/secrets/kubernetes.io/serviceaccount/namespace >> >> to find the namespace. >> >> >> On Tue, Jun 7, 2016 at 7:02 AM, Jorge Morales Pou >> wrote: >> >>> >>> http://stackoverflow.com/questions/31557932/how-to-get-the-namespace-from-inside-a-pod-in-openshift >>> >>> Jorge Morales >>> OpenShift by Red Hat >>> EMEA Field Product Manager >>> Product Marketing Manager >>> >>> On Tue, Jun 7, 2016 at 11:12 AM, Andrew Lee Rubinger >>> wrote: >>> >>>> Quick question in $subject. Was expecting some OPENSHIFT_ADDRESS or >>>> something environment variable passed into the Docker container when it >>>> started but I don't see it listed. >>>> >>>> S, >>>> ALR >>>> >>>> -- >>>> Red Hat Developer Programs Architecture >>>> @ALRubinger >>>> >>>> _______________________________________________ >>>> kontinuity-dev-public mailing list >>>> kontinuity-dev-public at redhat.com >>>> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >>>> >>>> >>> >>> _______________________________________________ >>> kontinuity-dev-public mailing list >>> kontinuity-dev-public at redhat.com >>> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >>> >>> >> >> >> -- >> Ben Parees | OpenShift >> >> > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From jenkins-kontinuity at redhat.com Tue Jun 7 21:00:41 2016 From: jenkins-kontinuity at redhat.com (jenkins-kontinuity at redhat.com) Date: Tue, 7 Jun 2016 21:00:41 +0000 (UTC) Subject: [kontinuity-dev-public] Jenkins build is still unstable: catapult #2 In-Reply-To: <116314017.0.1464634163117.JavaMail.root@jenkins-bm8by> References: <116314017.0.1464634163117.JavaMail.root@jenkins-bm8by> Message-ID: <326941301.1.1465333241060.JavaMail.root@jenkins-bm8by> See From jenkins-kontinuity at redhat.com Tue Jun 7 21:39:56 2016 From: jenkins-kontinuity at redhat.com (jenkins-kontinuity at redhat.com) Date: Tue, 7 Jun 2016 21:39:56 +0000 (UTC) Subject: [kontinuity-dev-public] Jenkins build is still unstable: catapult #3 In-Reply-To: <326941301.1.1465333241060.JavaMail.root@jenkins-bm8by> References: <326941301.1.1465333241060.JavaMail.root@jenkins-bm8by> Message-ID: <1813384367.2.1465335596474.JavaMail.root@jenkins-bm8by> See From jenkins-kontinuity at redhat.com Tue Jun 7 21:53:45 2016 From: jenkins-kontinuity at redhat.com (jenkins-kontinuity at redhat.com) Date: Tue, 7 Jun 2016 21:53:45 +0000 (UTC) Subject: [kontinuity-dev-public] Jenkins build is still unstable: catapult #4 In-Reply-To: <1813384367.2.1465335596474.JavaMail.root@jenkins-bm8by> References: <1813384367.2.1465335596474.JavaMail.root@jenkins-bm8by> Message-ID: <1486073471.3.1465336425197.JavaMail.root@jenkins-bm8by> See From jenkins-kontinuity at redhat.com Tue Jun 7 22:47:05 2016 From: jenkins-kontinuity at redhat.com (jenkins-kontinuity at redhat.com) Date: Tue, 7 Jun 2016 22:47:05 +0000 (UTC) Subject: [kontinuity-dev-public] Jenkins build is still unstable: catapult #5 In-Reply-To: <1486073471.3.1465336425197.JavaMail.root@jenkins-bm8by> References: <1486073471.3.1465336425197.JavaMail.root@jenkins-bm8by> Message-ID: <1719867026.4.1465339625399.JavaMail.root@jenkins-bm8by> See From jenkins-kontinuity at redhat.com Tue Jun 7 23:02:01 2016 From: jenkins-kontinuity at redhat.com (jenkins-kontinuity at redhat.com) Date: Tue, 7 Jun 2016 23:02:01 +0000 (UTC) Subject: [kontinuity-dev-public] Jenkins build is still unstable: catapult #6 In-Reply-To: <1719867026.4.1465339625399.JavaMail.root@jenkins-bm8by> References: <1719867026.4.1465339625399.JavaMail.root@jenkins-bm8by> Message-ID: <1354833518.5.1465340521282.JavaMail.root@jenkins-bm8by> See From alr at redhat.com Wed Jun 8 19:01:20 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Wed, 8 Jun 2016 15:01:20 -0400 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App Message-ID: Hi all: So I've been working on an OpenShift template to create the Catapult application. Users can fire up OpenShift, go to the console, add a new project "from file", upload the template, enter some required parameters, and it's there. What do I need to do to get this available "From Catalog" so users don't have to worry about finding/adding the template? S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From ablock at redhat.com Wed Jun 8 19:02:41 2016 From: ablock at redhat.com (Andrew Block) Date: Wed, 8 Jun 2016 15:02:41 -0400 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: References: Message-ID: You can add it to the ?openshift? project from a user with cluster-admin permissions oc create -f -n openshift? ? Andrew Block Red Hat Consulting 101 N. Wacker, Suite 150 Chicago, IL 60606 andrew.block at redhat.com?| m. (716) 870-2408 On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger (alr at redhat.com) wrote: Hi all: So I've been working on an OpenShift template to create the Catapult application.? Users can fire up OpenShift, go to the console, add a new project "from file", upload the template, enter some required parameters, and it's there. What do I need to do to get this available "From Catalog" so users don't have to worry about finding/adding the template? S, ALR -- Red Hat Developer Programs Architecture @ALRubinger _______________________________________________ kontinuity-dev-public mailing list kontinuity-dev-public at redhat.com https://www.redhat.com/mailman/listinfo/kontinuity-dev-public -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Wed Jun 8 19:07:46 2016 From: bparees at redhat.com (Ben Parees) Date: Wed, 8 Jun 2016 15:07:46 -0400 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: References: Message-ID: On Wed, Jun 8, 2016 at 3:02 PM, Andrew Block wrote: > You can add it to the ?openshift? project from a user with cluster-admin > permissions > > oc create -f -n openshift > > ?make sure it has the instantapp tag: https://github.com/openshift/origin/blob/master/examples/jenkins/jenkins-persistent-template.json#L10 ? > > > Andrew Block > Red Hat Consulting > 101 N. Wacker, Suite 150 > Chicago, IL 60606 > andrew.block at redhat.com | m. (716) 870-2408 > > On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger (alr at redhat.com) wrote: > > Hi all: > > So I've been working on an OpenShift template to create the Catapult > application. Users can fire up OpenShift, go to the console, add a new > project "from file", upload the template, enter some required parameters, > and it's there. > > What do I need to do to get this available "From Catalog" so users don't > have to worry about finding/adding the template? > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Wed Jun 8 19:13:07 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Wed, 8 Jun 2016 15:13:07 -0400 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: References: Message-ID: Ah, but since this is a local installation of OpenShift, the user would have to do it. I'm looking to do this for everyone so it appears alongside the other options (attached) on the distribution of Origin/ADB we make available at Summit. S, ALR On Wed, Jun 8, 2016 at 3:07 PM, Ben Parees wrote: > > > On Wed, Jun 8, 2016 at 3:02 PM, Andrew Block wrote: > >> You can add it to the ?openshift? project from a user with cluster-admin >> permissions >> >> oc create -f -n openshift >> >> > ?make sure it has the instantapp tag: > > https://github.com/openshift/origin/blob/master/examples/jenkins/jenkins-persistent-template.json#L10 > ? > > > >> >> >> Andrew Block >> Red Hat Consulting >> 101 N. Wacker, Suite 150 >> Chicago, IL 60606 >> andrew.block at redhat.com | m. (716) 870-2408 >> >> On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger (alr at redhat.com) >> wrote: >> >> Hi all: >> >> So I've been working on an OpenShift template to create the Catapult >> application. Users can fire up OpenShift, go to the console, add a new >> project "from file", upload the template, enter some required parameters, >> and it's there. >> >> What do I need to do to get this available "From Catalog" so users don't >> have to worry about finding/adding the template? >> >> S, >> ALR >> >> -- >> Red Hat Developer Programs Architecture >> @ALRubinger >> _______________________________________________ >> kontinuity-dev-public mailing list >> kontinuity-dev-public at redhat.com >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> >> >> _______________________________________________ >> kontinuity-dev-public mailing list >> kontinuity-dev-public at redhat.com >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> >> > > > -- > Ben Parees | OpenShift > > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: screenshot2.png Type: image/png Size: 53292 bytes Desc: not available URL: From bparees at redhat.com Wed Jun 8 19:16:30 2016 From: bparees at redhat.com (Ben Parees) Date: Wed, 8 Jun 2016 15:16:30 -0400 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: References: Message-ID: On Wed, Jun 8, 2016 at 3:13 PM, Andrew Lee Rubinger wrote: > Ah, but since this is a local installation of OpenShift, the user would > have to do it. > > I'm looking to do this for everyone so it appears alongside the other > options (attached) on the distribution of Origin/ADB we make available at > Summit. > ?well, first off everything you screenshotted there is a builder image, not a template or instantapp. ? the set of templates that the installer registers w/ the openshift namespace is determined by the ansible installer (or other mechanisms depending on where your cluster came from) so we'd have to ship your template which i doubt we want to do. If you want it registered in the ADB VM, then you can do that as part of the VM setup. > > S, > ALR > > On Wed, Jun 8, 2016 at 3:07 PM, Ben Parees wrote: > >> >> >> On Wed, Jun 8, 2016 at 3:02 PM, Andrew Block wrote: >> >>> You can add it to the ?openshift? project from a user with cluster-admin >>> permissions >>> >>> oc create -f -n openshift >>> >>> >> ?make sure it has the instantapp tag: >> >> https://github.com/openshift/origin/blob/master/examples/jenkins/jenkins-persistent-template.json#L10 >> ? >> >> >> >>> >>> >>> Andrew Block >>> Red Hat Consulting >>> 101 N. Wacker, Suite 150 >>> Chicago, IL 60606 >>> andrew.block at redhat.com | m. (716) 870-2408 >>> >>> On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger (alr at redhat.com) >>> wrote: >>> >>> Hi all: >>> >>> So I've been working on an OpenShift template to create the Catapult >>> application. Users can fire up OpenShift, go to the console, add a new >>> project "from file", upload the template, enter some required parameters, >>> and it's there. >>> >>> What do I need to do to get this available "From Catalog" so users don't >>> have to worry about finding/adding the template? >>> >>> S, >>> ALR >>> >>> -- >>> Red Hat Developer Programs Architecture >>> @ALRubinger >>> _______________________________________________ >>> kontinuity-dev-public mailing list >>> kontinuity-dev-public at redhat.com >>> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >>> >>> >>> _______________________________________________ >>> kontinuity-dev-public mailing list >>> kontinuity-dev-public at redhat.com >>> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >>> >>> >> >> >> -- >> Ben Parees | OpenShift >> >> > > > -- > Red Hat Developer Programs Architecture > @ALRubinger > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Wed Jun 8 19:18:42 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Wed, 8 Jun 2016 15:18:42 -0400 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: References: Message-ID: On Wed, Jun 8, 2016 at 3:16 PM, Ben Parees wrote: > > > On Wed, Jun 8, 2016 at 3:13 PM, Andrew Lee Rubinger > wrote: > >> Ah, but since this is a local installation of OpenShift, the user would >> have to do it. >> >> I'm looking to do this for everyone so it appears alongside the other >> options (attached) on the distribution of Origin/ADB we make available at >> Summit. >> > > > ?well, first off everything you screenshotted there is a builder image, > not a template or instantapp. > ? > the set of templates that the installer registers w/ the openshift > namespace is determined by the ansible installer (or other mechanisms > depending on where your cluster came from) so we'd have to ship your > template which i doubt we want to do. If you want it registered in the ADB > VM, then you can do that as part of the VM setup. > Very helpful, thank you. :) > > > > > >> >> S, >> ALR >> >> On Wed, Jun 8, 2016 at 3:07 PM, Ben Parees wrote: >> >>> >>> >>> On Wed, Jun 8, 2016 at 3:02 PM, Andrew Block wrote: >>> >>>> You can add it to the ?openshift? project from a user with >>>> cluster-admin permissions >>>> >>>> oc create -f -n openshift >>>> >>>> >>> ?make sure it has the instantapp tag: >>> >>> https://github.com/openshift/origin/blob/master/examples/jenkins/jenkins-persistent-template.json#L10 >>> ? >>> >>> >>> >>>> >>>> >>>> Andrew Block >>>> Red Hat Consulting >>>> 101 N. Wacker, Suite 150 >>>> Chicago, IL 60606 >>>> andrew.block at redhat.com | m. (716) 870-2408 >>>> >>>> On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger (alr at redhat.com) >>>> wrote: >>>> >>>> Hi all: >>>> >>>> So I've been working on an OpenShift template to create the Catapult >>>> application. Users can fire up OpenShift, go to the console, add a new >>>> project "from file", upload the template, enter some required parameters, >>>> and it's there. >>>> >>>> What do I need to do to get this available "From Catalog" so users >>>> don't have to worry about finding/adding the template? >>>> >>>> S, >>>> ALR >>>> >>>> -- >>>> Red Hat Developer Programs Architecture >>>> @ALRubinger >>>> _______________________________________________ >>>> kontinuity-dev-public mailing list >>>> kontinuity-dev-public at redhat.com >>>> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >>>> >>>> >>>> _______________________________________________ >>>> kontinuity-dev-public mailing list >>>> kontinuity-dev-public at redhat.com >>>> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >>>> >>>> >>> >>> >>> -- >>> Ben Parees | OpenShift >>> >>> >> >> >> -- >> Red Hat Developer Programs Architecture >> @ALRubinger >> > > > > -- > Ben Parees | OpenShift > > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Thu Jun 9 08:07:19 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Thu, 09 Jun 2016 10:07:19 +0200 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: References: Message-ID: <1465459639.4017.4.camel@redhat.com> On St, 2016-06-08 at 15:18 -0400, Andrew Lee Rubinger wrote: > > > On Wed, Jun 8, 2016 at 3:16 PM, Ben Parees > wrote: > > > > > > On Wed, Jun 8, 2016 at 3:13 PM, Andrew Lee Rubinger > > wrote: > > > Ah, but since this is a local installation of OpenShift, the user > > > would have to do it. > > > > > > I'm looking to do this for everyone so it appears alongside the > > > other options (attached) on the distribution of Origin/ADB we > > > make available at Summit. > > > > > > > well, first off everything you screenshotted there is a builder > > image, not a template or instantapp. > > > > the set of templates that the installer registers w/ the openshift > > namespace is determined by the ansible installer (or other > > mechanisms depending on where your cluster came from) so we'd have > > to ship your template which i doubt we want to do.? If you want it > > registered in the ADB VM, then you can do that as part of the VM > > setup. > > > Very helpful, thank you. :) 1. If you want it only in ADB just drop it into?https://github.com/proj ectatomic/adb-utils/tree/master/services/openshift/templates/adb 2. If you want it in both ADB and CDK drop it into?https://github.com/p rojectatomic/adb-utils/tree/master/services/openshift/templates/common And, of course send an PR :) (Release is coming soon!?https://github.com/projectatomic/adb-utils/iss ues/140) > ? > > > > > > > > ? > > > > > > S, > > > ALR > > > > > > On Wed, Jun 8, 2016 at 3:07 PM, Ben Parees > > > wrote: > > > > > > > > > > > > On Wed, Jun 8, 2016 at 3:02 PM, Andrew Block > > > > wrote: > > > > > You can add it to the ?openshift? project from a user with > > > > > cluster-admin permissions > > > > > > > > > > oc create -f -n openshift? > > > > > > > > > > > > > > make sure it has the instantapp tag: > > > > https://github.com/openshift/origin/blob/master/examples/jenkin > > > > s/jenkins-persistent-template.json#L10 > > > > > > > > > > > > ? > > > > > > > > > > ? > > > > > Andrew Block > > > > > Red Hat Consulting > > > > > 101 N. Wacker, Suite 150 > > > > > Chicago, IL 60606 > > > > > andrew.block at redhat.com?| m. (716) 870-2408 > > > > > > > > > > On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger > > > > > (alr at redhat.com) wrote: > > > > > > Hi all: > > > > > > > > > > > > So I've been working on an OpenShift template to create the > > > > > > Catapult application.? Users can fire up OpenShift, go to > > > > > > the console, add a new project "from file", upload the > > > > > > template, enter some required parameters, and it's there. > > > > > > > > > > > > What do I need to do to get this available "From Catalog" > > > > > > so users don't have to worry about finding/adding the > > > > > > template? > > > > > > > > > > > > S, > > > > > > ALR > > > > > > > > > > > > -- > > > > > > Red Hat Developer Programs Architecture > > > > > > @ALRubinger > > > > > > _______________________________________________? > > > > > > kontinuity-dev-public mailing list? > > > > > > kontinuity-dev-public at redhat.com? > > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev- > > > > > > public? > > > > > _______________________________________________ > > > > > kontinuity-dev-public mailing list > > > > > kontinuity-dev-public at redhat.com > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > > > > > > > > > > > > > > --? > > > > Ben Parees | OpenShift > > > > > > > > > > > > > > > > > --? > > > Red Hat Developer Programs Architecture > > > @ALRubinger > > > > > > > > > _______________________________________________ > > kontinuity-dev-public mailing list > > kontinuity-dev-public at redhat.com > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Thu Jun 9 12:28:06 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Thu, 9 Jun 2016 08:28:06 -0400 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: <1465459639.4017.4.camel@redhat.com> References: <1465459639.4017.4.camel@redhat.com> Message-ID: On Thu, Jun 9, 2016 at 4:07 AM, Tomas Nozicka wrote: > On St, 2016-06-08 at 15:18 -0400, Andrew Lee Rubinger wrote: > > > > On Wed, Jun 8, 2016 at 3:16 PM, Ben Parees wrote: > > > > On Wed, Jun 8, 2016 at 3:13 PM, Andrew Lee Rubinger > wrote: > > Ah, but since this is a local installation of OpenShift, the user would > have to do it. > > I'm looking to do this for everyone so it appears alongside the other > options (attached) on the distribution of Origin/ADB we make available at > Summit. > > > well, first off everything you screenshotted there is a builder image, not > a template or instantapp. > > the set of templates that the installer registers w/ the openshift > namespace is determined by the ansible installer (or other mechanisms > depending on where your cluster came from) so we'd have to ship your > template which i doubt we want to do. If you want it registered in the ADB > VM, then you can do that as part of the VM setup. > > Very helpful, thank you. :) > > 1. If you want it only in ADB just drop it into > https://github.com/projectatomic/adb-utils/tree/master/services/openshift/templates/adb > 2. If you want it in both ADB and CDK drop it into > https://github.com/projectatomic/adb-utils/tree/master/services/openshift/templates/common > > And, of course send an PR :) > (Release is coming soon! > https://github.com/projectatomic/adb-utils/issues/140) > Thank you; due date? In the office today for meetings so my dev time is maybe on hold until Friday. S, ALR > > > > > > > > > > > S, > ALR > > On Wed, Jun 8, 2016 at 3:07 PM, Ben Parees wrote: > > > > On Wed, Jun 8, 2016 at 3:02 PM, Andrew Block wrote: > > You can add it to the ?openshift? project from a user with cluster-admin > permissions > > oc create -f -n openshift > > > make sure it has the instantapp tag: > > https://github.com/openshift/origin/blob/master/examples/jenkins/jenkins-persistent-template.json#L10 > > > > > > > Andrew Block > Red Hat Consulting > 101 N. Wacker, Suite 150 > Chicago, IL 60606 > andrew.block at redhat.com | m. (716) 870-2408 > > On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger (alr at redhat.com) wrote: > > Hi all: > > So I've been working on an OpenShift template to create the Catapult > application. Users can fire up OpenShift, go to the console, add a new > project "from file", upload the template, enter some required parameters, > and it's there. > > What do I need to do to get this available "From Catalog" so users don't > have to worry about finding/adding the template? > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > -- > Ben Parees | OpenShift > > > > > -- > Red Hat Developer Programs Architecture > @ALRubinger > > > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Thu Jun 9 17:08:30 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Thu, 09 Jun 2016 19:08:30 +0200 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: References: <1465459639.4017.4.camel@redhat.com> Message-ID: <1465492110.8745.3.camel@redhat.com> On ?t, 2016-06-09 at 08:28 -0400, Andrew Lee Rubinger wrote: > > > On Thu, Jun 9, 2016 at 4:07 AM, Tomas Nozicka > wrote: > > On St, 2016-06-08 at 15:18 -0400, Andrew Lee Rubinger wrote: > > > > > > > > > On Wed, Jun 8, 2016 at 3:16 PM, Ben Parees > > > wrote: > > > > > > > > > > > > On Wed, Jun 8, 2016 at 3:13 PM, Andrew Lee Rubinger > > > .com> wrote: > > > > > Ah, but since this is a local installation of OpenShift, the > > > > > user would have to do it. > > > > > > > > > > I'm looking to do this for everyone so it appears alongside > > > > > the other options (attached) on the distribution of > > > > > Origin/ADB we make available at Summit. > > > > > > > > > well, first off everything you screenshotted there is a builder > > > > image, not a template or instantapp. > > > > > > > > the set of templates that the installer registers w/ the > > > > openshift namespace is determined by the ansible installer (or > > > > other mechanisms depending on where your cluster came from) so > > > > we'd have to ship your template which i doubt we want to do.? > > > > If you want it registered in the ADB VM, then you can do that > > > > as part of the VM setup. > > > > > > > Very helpful, thank you. :) > > 1. If you want it only in ADB just drop it into?https://github.com/ > > projectatomic/adb- > > utils/tree/master/services/openshift/templates/adb > > 2. If you want it in both ADB and CDK drop it into?https://github.c > > om/projectatomic/adb- > > utils/tree/master/services/openshift/templates/common > > > > And, of course send an PR :) > > (Release is coming soon!?https://github.com/projectatomic/adb-utils > > /issues/140) > > > Thank you; due date?? In the office today for meetings so my dev time > is maybe on hold until Friday. Release should be tomorrow, but you can make it to the next one. Although I am pretty certain that we can install you template (instantapp) from Vagrantfile though ? oc create -f without the need for any release. > > S, > ALR > ? > > > > > > > ? > > > > > > > > > > > > > > > > ? > > > > > > > > > > S, > > > > > ALR > > > > > > > > > > On Wed, Jun 8, 2016 at 3:07 PM, Ben Parees > > > > m> wrote: > > > > > > > > > > > > > > > > > > On Wed, Jun 8, 2016 at 3:02 PM, Andrew Block > > > > > .com> wrote: > > > > > > > You can add it to the ?openshift? project from a user > > > > > > > with cluster-admin permissions > > > > > > > > > > > > > > oc create -f -n openshift? > > > > > > > > > > > > > > > > > > > > make sure it has the instantapp tag: > > > > > > https://github.com/openshift/origin/blob/master/examples/je > > > > > > nkins/jenkins-persistent-template.json#L10 > > > > > > > > > > > > > > > > > > ? > > > > > > > > > > > > > > ? > > > > > > > Andrew Block > > > > > > > Red Hat Consulting > > > > > > > 101 N. Wacker, Suite 150 > > > > > > > Chicago, IL 60606 > > > > > > > andrew.block at redhat.com?| m. (716) 870-2408 > > > > > > > > > > > > > > On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger (alr at r > > > > > > > edhat.com) wrote: > > > > > > > > Hi all: > > > > > > > > > > > > > > > > So I've been working on an OpenShift template to create > > > > > > > > the Catapult application.? Users can fire up OpenShift, > > > > > > > > go to the console, add a new project "from file", > > > > > > > > upload the template, enter some required parameters, > > > > > > > > and it's there. > > > > > > > > > > > > > > > > What do I need to do to get this available "From > > > > > > > > Catalog" so users don't have to worry about > > > > > > > > finding/adding the template? > > > > > > > > > > > > > > > > S, > > > > > > > > ALR > > > > > > > > > > > > > > > > -- > > > > > > > > Red Hat Developer Programs Architecture > > > > > > > > @ALRubinger > > > > > > > > _______________________________________________? > > > > > > > > kontinuity-dev-public mailing list? > > > > > > > > kontinuity-dev-public at redhat.com? > > > > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev- > > > > > > > > public? > > > > > > > _______________________________________________ > > > > > > > kontinuity-dev-public mailing list > > > > > > > kontinuity-dev-public at redhat.com > > > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-pu > > > > > > > blic > > > > > > > > > > > > > > > > > > > --? > > > > > > Ben Parees | OpenShift > > > > > > > > > > > > > > > > > > > > > > --? > > > > > Red Hat Developer Programs Architecture > > > > > @ALRubinger > > > > > > > > > > > > > _______________________________________________ > > > > kontinuity-dev-public mailing list > > > > kontinuity-dev-public at redhat.com > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > > > > --? > Red Hat Developer Programs Architecture > @ALRubinger From rmartine at redhat.com Thu Jun 9 18:21:20 2016 From: rmartine at redhat.com (Ricardo Martinelli de Oliveira) Date: Thu, 9 Jun 2016 15:21:20 -0300 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: <1465492110.8745.3.camel@redhat.com> References: <1465459639.4017.4.camel@redhat.com> <1465492110.8745.3.camel@redhat.com> Message-ID: ALR, Gimme all you have and I can move forward with it. On Thu, Jun 9, 2016 at 2:08 PM, Tomas Nozicka wrote: > On ?t, 2016-06-09 at 08:28 -0400, Andrew Lee Rubinger wrote: > > > > > > On Thu, Jun 9, 2016 at 4:07 AM, Tomas Nozicka > > wrote: > > > On St, 2016-06-08 at 15:18 -0400, Andrew Lee Rubinger wrote: > > > > > > > > > > > > On Wed, Jun 8, 2016 at 3:16 PM, Ben Parees > > > > wrote: > > > > > > > > > > > > > > > On Wed, Jun 8, 2016 at 3:13 PM, Andrew Lee Rubinger > > > > .com> wrote: > > > > > > Ah, but since this is a local installation of OpenShift, the > > > > > > user would have to do it. > > > > > > > > > > > > I'm looking to do this for everyone so it appears alongside > > > > > > the other options (attached) on the distribution of > > > > > > Origin/ADB we make available at Summit. > > > > > > > > > > > well, first off everything you screenshotted there is a builder > > > > > image, not a template or instantapp. > > > > > > > > > > the set of templates that the installer registers w/ the > > > > > openshift namespace is determined by the ansible installer (or > > > > > other mechanisms depending on where your cluster came from) so > > > > > we'd have to ship your template which i doubt we want to do. > > > > > If you want it registered in the ADB VM, then you can do that > > > > > as part of the VM setup. > > > > > > > > > Very helpful, thank you. :) > > > 1. If you want it only in ADB just drop it into https://github.com/ > > > projectatomic/adb- > > > utils/tree/master/services/openshift/templates/adb > > > 2. If you want it in both ADB and CDK drop it into https://github.c > > > om/projectatomic/adb- > > > utils/tree/master/services/openshift/templates/common > > > > > > And, of course send an PR :) > > > (Release is coming soon! https://github.com/projectatomic/adb-utils > > > /issues/140) > > > > > Thank you; due date? In the office today for meetings so my dev time > > is maybe on hold until Friday. > Release should be tomorrow, but you can make it to the next one. > Although I am pretty certain that we can install you template > (instantapp) from Vagrantfile though > oc create -f > without the need for any release. > > > > > S, > > ALR > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > S, > > > > > > ALR > > > > > > > > > > > > On Wed, Jun 8, 2016 at 3:07 PM, Ben Parees > > > > > m> wrote: > > > > > > > > > > > > > > > > > > > > > On Wed, Jun 8, 2016 at 3:02 PM, Andrew Block > > > > > > .com> wrote: > > > > > > > > You can add it to the ?openshift? project from a user > > > > > > > > with cluster-admin permissions > > > > > > > > > > > > > > > > oc create -f -n openshift > > > > > > > > > > > > > > > > > > > > > > > make sure it has the instantapp tag: > > > > > > > https://github.com/openshift/origin/blob/master/examples/je > > > > > > > nkins/jenkins-persistent-template.json#L10 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Andrew Block > > > > > > > > Red Hat Consulting > > > > > > > > 101 N. Wacker, Suite 150 > > > > > > > > Chicago, IL 60606 > > > > > > > > andrew.block at redhat.com | m. (716) 870-2408 > > > > > > > > > > > > > > > > On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger (alr at r > > > > > > > > edhat.com) wrote: > > > > > > > > > Hi all: > > > > > > > > > > > > > > > > > > So I've been working on an OpenShift template to create > > > > > > > > > the Catapult application. Users can fire up OpenShift, > > > > > > > > > go to the console, add a new project "from file", > > > > > > > > > upload the template, enter some required parameters, > > > > > > > > > and it's there. > > > > > > > > > > > > > > > > > > What do I need to do to get this available "From > > > > > > > > > Catalog" so users don't have to worry about > > > > > > > > > finding/adding the template? > > > > > > > > > > > > > > > > > > S, > > > > > > > > > ALR > > > > > > > > > > > > > > > > > > -- > > > > > > > > > Red Hat Developer Programs Architecture > > > > > > > > > @ALRubinger > > > > > > > > > _______________________________________________ > > > > > > > > > kontinuity-dev-public mailing list > > > > > > > > > kontinuity-dev-public at redhat.com > > > > > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev- > > > > > > > > > public > > > > > > > > _______________________________________________ > > > > > > > > kontinuity-dev-public mailing list > > > > > > > > kontinuity-dev-public at redhat.com > > > > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-pu > > > > > > > > blic > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > Ben Parees | OpenShift > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > Red Hat Developer Programs Architecture > > > > > > @ALRubinger > > > > > > > > > > > > > > > > _______________________________________________ > > > > > kontinuity-dev-public mailing list > > > > > kontinuity-dev-public at redhat.com > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > > > > > > > > > > -- > > Red Hat Developer Programs Architecture > > @ALRubinger > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Thu Jun 9 18:52:45 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Thu, 9 Jun 2016 14:52:45 -0400 Subject: [kontinuity-dev-public] Making Catapult into an OpenShift Instant App In-Reply-To: References: <1465459639.4017.4.camel@redhat.com> <1465492110.8745.3.camel@redhat.com> Message-ID: So far, the commit linked here: https://github.com/redhat-kontinuity/catapult/issues/114 ...but I haven't merged it upstream or had Ben or someone sanity-check it. So far it just kinda works "for me". I am planning on further reviewing it tonight and putting it upstream for others to consume and try. Thanks all! S, ALR On Thu, Jun 9, 2016 at 2:21 PM, Ricardo Martinelli de Oliveira < rmartine at redhat.com> wrote: > ALR, > > Gimme all you have and I can move forward with it. > > On Thu, Jun 9, 2016 at 2:08 PM, Tomas Nozicka wrote: > >> On ?t, 2016-06-09 at 08:28 -0400, Andrew Lee Rubinger wrote: >> > >> > >> > On Thu, Jun 9, 2016 at 4:07 AM, Tomas Nozicka >> > wrote: >> > > On St, 2016-06-08 at 15:18 -0400, Andrew Lee Rubinger wrote: >> > > > >> > > > >> > > > On Wed, Jun 8, 2016 at 3:16 PM, Ben Parees >> > > > wrote: >> > > > > >> > > > > >> > > > > On Wed, Jun 8, 2016 at 3:13 PM, Andrew Lee Rubinger > > > > > .com> wrote: >> > > > > > Ah, but since this is a local installation of OpenShift, the >> > > > > > user would have to do it. >> > > > > > >> > > > > > I'm looking to do this for everyone so it appears alongside >> > > > > > the other options (attached) on the distribution of >> > > > > > Origin/ADB we make available at Summit. >> > > > > > >> > > > > well, first off everything you screenshotted there is a builder >> > > > > image, not a template or instantapp. >> > > > > >> > > > > the set of templates that the installer registers w/ the >> > > > > openshift namespace is determined by the ansible installer (or >> > > > > other mechanisms depending on where your cluster came from) so >> > > > > we'd have to ship your template which i doubt we want to do. >> > > > > If you want it registered in the ADB VM, then you can do that >> > > > > as part of the VM setup. >> > > > > >> > > > Very helpful, thank you. :) >> > > 1. If you want it only in ADB just drop it into https://github.com/ >> > > projectatomic/adb- >> > > utils/tree/master/services/openshift/templates/adb >> > > 2. If you want it in both ADB and CDK drop it into https://github.c >> > > om/projectatomic/adb- >> > > utils/tree/master/services/openshift/templates/common >> > > >> > > And, of course send an PR :) >> > > (Release is coming soon! https://github.com/projectatomic/adb-utils >> > > /issues/140) >> > > >> > Thank you; due date? In the office today for meetings so my dev time >> > is maybe on hold until Friday. >> Release should be tomorrow, but you can make it to the next one. >> Although I am pretty certain that we can install you template >> (instantapp) from Vagrantfile though >> oc create -f >> without the need for any release. >> >> > >> > S, >> > ALR >> > >> > > >> > > >> > > > >> > > > > >> > > > > >> > > > > >> > > > > >> > > > > > >> > > > > > S, >> > > > > > ALR >> > > > > > >> > > > > > On Wed, Jun 8, 2016 at 3:07 PM, Ben Parees > > > > > > m> wrote: >> > > > > > > >> > > > > > > >> > > > > > > On Wed, Jun 8, 2016 at 3:02 PM, Andrew Block > > > > > > > .com> wrote: >> > > > > > > > You can add it to the ?openshift? project from a user >> > > > > > > > with cluster-admin permissions >> > > > > > > > >> > > > > > > > oc create -f -n openshift >> > > > > > > > >> > > > > > > > >> > > > > > > make sure it has the instantapp tag: >> > > > > > > https://github.com/openshift/origin/blob/master/examples/je >> > > > > > > nkins/jenkins-persistent-template.json#L10 >> > > > > > > >> > > > > > > >> > > > > > > >> > > > > > > > >> > > > > > > > >> > > > > > > > Andrew Block >> > > > > > > > Red Hat Consulting >> > > > > > > > 101 N. Wacker, Suite 150 >> > > > > > > > Chicago, IL 60606 >> > > > > > > > andrew.block at redhat.com | m. (716) 870-2408 >> > > > > > > > >> > > > > > > > On June 8, 2016 at 3:01:51 PM, Andrew Lee Rubinger (alr at r >> > > > > > > > edhat.com) wrote: >> > > > > > > > > Hi all: >> > > > > > > > > >> > > > > > > > > So I've been working on an OpenShift template to create >> > > > > > > > > the Catapult application. Users can fire up OpenShift, >> > > > > > > > > go to the console, add a new project "from file", >> > > > > > > > > upload the template, enter some required parameters, >> > > > > > > > > and it's there. >> > > > > > > > > >> > > > > > > > > What do I need to do to get this available "From >> > > > > > > > > Catalog" so users don't have to worry about >> > > > > > > > > finding/adding the template? >> > > > > > > > > >> > > > > > > > > S, >> > > > > > > > > ALR >> > > > > > > > > >> > > > > > > > > -- >> > > > > > > > > Red Hat Developer Programs Architecture >> > > > > > > > > @ALRubinger >> > > > > > > > > _______________________________________________ >> > > > > > > > > kontinuity-dev-public mailing list >> > > > > > > > > kontinuity-dev-public at redhat.com >> > > > > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev- >> > > > > > > > > public >> > > > > > > > _______________________________________________ >> > > > > > > > kontinuity-dev-public mailing list >> > > > > > > > kontinuity-dev-public at redhat.com >> > > > > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-pu >> > > > > > > > blic >> > > > > > > > >> > > > > > > >> > > > > > > -- >> > > > > > > Ben Parees | OpenShift >> > > > > > > >> > > > > > > >> > > > > > >> > > > > > -- >> > > > > > Red Hat Developer Programs Architecture >> > > > > > @ALRubinger >> > > > > > >> > > > > >> > > > > _______________________________________________ >> > > > > kontinuity-dev-public mailing list >> > > > > kontinuity-dev-public at redhat.com >> > > > > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> > > > > >> > >> > >> > -- >> > Red Hat Developer Programs Architecture >> > @ALRubinger >> >> _______________________________________________ >> kontinuity-dev-public mailing list >> kontinuity-dev-public at redhat.com >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> > > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Fri Jun 10 01:26:01 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Thu, 9 Jun 2016 21:26:01 -0400 Subject: [kontinuity-dev-public] Request for OS Template file review Message-ID: Hi all: Anyone familiar with OS template files that can review this? https://github.com/redhat-kontinuity/catapult/commit/b9404f83771343838e8d32fb8fbc53aa5044bafe To use: create a new project then process this template either through the console "From file..." or using oc. In particular, why define the container config and env vars twice (line 105 and 178)? Any other extraneous stuff from the "oc export" command we can/should remove from here? S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From ablock at redhat.com Fri Jun 10 01:30:45 2016 From: ablock at redhat.com (Andrew Block) Date: Thu, 9 Jun 2016 21:30:45 -0400 Subject: [kontinuity-dev-public] Request for OS Template file review In-Reply-To: References: Message-ID: I am not a huge fan of the export as template command as it generates additional metadata that is not as useful in templates. These include: lastTransactionTime observedGeneration status fields In addition, I would recommend migrating from a ReplicationController to a DeploymentController as it aligns with the OpenShift model as well as making the majority of the environment variables in the container as template parameters in case they would need to be overridden. - Andy ? Andrew Block Red Hat Consulting 101 N. Wacker, Suite 150 Chicago, IL 60606 andrew.block at redhat.com?| m. (716) 870-2408 On June 9, 2016 at 9:26:33 PM, Andrew Lee Rubinger (alr at redhat.com) wrote: Hi all: Anyone familiar with OS template files that can review this? ??https://github.com/redhat-kontinuity/catapult/commit/b9404f83771343838e8d32fb8fbc53aa5044bafe To use: create a new project then process this template either through the console "From file..." or using oc. In particular, why define the container config and env vars twice (line 105 and 178)? Any other extraneous stuff from the "oc export" command we can/should remove from here? S, ALR -- Red Hat Developer Programs Architecture @ALRubinger _______________________________________________ kontinuity-dev-public mailing list kontinuity-dev-public at redhat.com https://www.redhat.com/mailman/listinfo/kontinuity-dev-public -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Fri Jun 10 14:11:35 2016 From: bparees at redhat.com (Ben Parees) Date: Fri, 10 Jun 2016 10:11:35 -0400 Subject: [kontinuity-dev-public] Request for OS Template file review In-Reply-To: References: Message-ID: I made a few comments in the commit, but one general comment is that you probably want to parameterize more of the names/labels/selector values so people can avoid collisions. see: https://github.com/openshift/origin/blob/master/examples/quickstarts/nodejs-mongodb.json#L248 https://github.com/openshift/origin/blob/master/examples/quickstarts/nodejs-mongodb.json#L226 https://github.com/openshift/origin/blob/master/examples/quickstarts/nodejs-mongodb.json#L283-L285 https://github.com/openshift/origin/blob/master/examples/quickstarts/nodejs-mongodb.json#L239-L241 On Thu, Jun 9, 2016 at 9:26 PM, Andrew Lee Rubinger wrote: > Hi all: > > Anyone familiar with OS template files that can review this? > > > https://github.com/redhat-kontinuity/catapult/commit/b9404f83771343838e8d32fb8fbc53aa5044bafe > > To use: create a new project then process this template either through the > console "From file..." or using oc. > > In particular, why define the container config and env vars twice (line > 105 and 178)? > > Any other extraneous stuff from the "oc export" command we can/should > remove from here? > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Wed Jun 15 09:23:18 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Wed, 15 Jun 2016 05:23:18 -0400 Subject: [kontinuity-dev-public] Breaking Changes in Catapult: CI Will Fail Message-ID: Primarily for AndrewB and Dharmit: @see: https://github.com/redhat-kontinuity/catapult/commit/a5895b924e20e64591508db935a73b56803fbd28 We merged in Xavier's awesome and feared #97 fixes tonight, and as part of that we need to do some manual work w/ OpenShift to add the Jenkins Template. The CI jobs will now fail until they're equipped to handle the same; is it clear enough from my diffs to the README above what needs to be done? S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Thu Jun 16 00:07:28 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Wed, 15 Jun 2016 20:07:28 -0400 Subject: [kontinuity-dev-public] The User Experience As It Stands Message-ID: Hi all: OK, so at this point I think we have enough integration work done and user experience instructions set forth that I hope we're all starting to look at this from the same perspective. Apologies to any non-RHTers; I'm not sure how to share this outside the org at the moment: https://docs.google.com/a/redhat.com/document/d/1FtBt8XwZ237Mko6zEpkTOuFFhHfXM9THerjh702XA0c/edit?usp=sharing If everyone can center on that as the "thing to build out" and do dev against, that'd be wonderful. Orange are areas we want to improve but don't necessarily block us, and red is blocking us from saying "we're done here". Following those instructions, we get nicely down to observing the OpenShift project overview page for the application Catapult has just flung. Looks a little like this at the moment: https://imagebin.ca/v/2kniMSCFgVhS This is where I get a bit lost. :) We have 3 services: * Jenkins * jenkins-jnlp * kitchensink-html5-mobile (the example app in question) The example app then starts a build using s2i and *almost* completes: https://gist.github.com/ALRubinger/fd81272e2f8eaf0e07b8f4b61c8dab42 The Jenkins instance has 2 projects, "pipeline" and "pipeline-example", and I'm unsure how they relate to the example app we want to build/deploy. I have a sneaking suspicion that I've been informed that many of you are working on reconciling how this works together for your feature areas, but I'm not yet grokking the whole picture. :) 1) What's the intended role of the Jenkins service here? Shouldn't it be handling the pipeline build? 2) Why is the build being kicked off in the example service itself? Is that what we want? 3) Who's working on these pieces and are we all clear on what the experience should be? I know Tomas and Xavier are syncing in the morning about the template file approach to take, so hoping that discussion lends some clarity to the questions above. :) In the meantime I'll be working on the yellow items. S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Thu Jun 16 02:00:42 2016 From: bparees at redhat.com (Ben Parees) Date: Wed, 15 Jun 2016 19:00:42 -0700 Subject: [kontinuity-dev-public] The User Experience As It Stands In-Reply-To: References: Message-ID: On Wed, Jun 15, 2016 at 5:07 PM, Andrew Lee Rubinger wrote: > Hi all: > > OK, so at this point I think we have enough integration work done and user > experience instructions set forth that I hope we're all starting to look at > this from the same perspective. > > Apologies to any non-RHTers; I'm not sure how to share this outside the > org at the moment: > > > https://docs.google.com/a/redhat.com/document/d/1FtBt8XwZ237Mko6zEpkTOuFFhHfXM9THerjh702XA0c/edit?usp=sharing > > If everyone can center on that as the "thing to build out" and do dev > against, that'd be wonderful. > > Orange are areas we want to improve but don't necessarily block us, and > red is blocking us from saying "we're done here". > > Following those instructions, we get nicely down to observing the > OpenShift project overview page for the application Catapult has just > flung. Looks a little like this at the moment: > > https://imagebin.ca/v/2kniMSCFgVhS > > This is where I get a bit lost. :) > > We have 3 services: > > * Jenkins > * jenkins-jnlp > * kitchensink-html5-mobile (the example app in question) > > The example app then starts a build using s2i and *almost* completes: > > https://gist.github.com/ALRubinger/fd81272e2f8eaf0e07b8f4b61c8dab42 > ?that's indicative of a problem w/ your cluster. i assume even if you created a basic s2i buildconfig and kicked it off (no jenkins/pipelines involved) you'd hit the same issue, so you'll need to dig into why your registry appears to be inaccessible in your cluster. > > The Jenkins instance has 2 projects, "pipeline" and "pipeline-example", > and I'm unsure how they relate to the example app we want to build/deploy. > ?pipeline-example is built into the jenkins image, it's part of the basic config we include in the image. we may get rid of it soon since it's causing confusion, but it was intended as an out of the box "here's a pipeline sample job": https://github.com/fabric8io/openshift-jenkins-s2i-config/tree/master/configuration/jobs/pipeline-example "pipeline" is, i assume, the name of your buildconfig that you're actually interested in, that gets created for you when we see the buildconfig in your openshift project. (I don't know for sure, the one we provide in the tutorial is named "sample-pipeline", so i don't know where "pipeline" came from in your case, but one way or another, you created it and it's the one you care about) > > I have a sneaking suspicion that I've been informed that many of you are > working on reconciling how this works together for your feature areas, but > I'm not yet grokking the whole picture. :) > > 1) What's the intended role of the Jenkins service here? Shouldn't it be > handling the pipeline build? > ?the jenkins service provides access the the jenkins web console? and apis. > 2) Why is the build being kicked off in the example service itself? Is > that what we want? > ?not sure what you mean by a build being kicked off in the example service.? The normal flow is: 1) create an openshift buildconfig w/ strategy type pipeline 2) a jenkins service+deployment gets created for you 3) a new jenkins job gets created in the jenkins server that corresponds to the buildconfig from (1) 4) use openshift to start a build of the buildconfig from (1) 5) that creates an openshift build object based on your buildconfig 6) the fabric sync plugin running in the jenkins server from (2) detects the new build object 7) the fabric sync plugin running in the jenkins server starts the jenkins job from (3) 8) the jenkins pipeline job does whatever you defined it to do (which may or may not include launching slave pods on openshift) 9) the fabric sync plugin updates the openshift build object from (5) with the ongoing state of the pipeline execution (what stage it's in, success/failure, etc). > 3) Who's working on these pieces and are we all clear on what the > experience should be? > ?my team, jessica's team, the fabric team, your team. I feel pretty clear about all the bits on the openshift/fabric/jenkins side. I haven't worried about the Catapult layer since that's above those things as i understand it. > > I know Tomas and Xavier are syncing in the morning about the template file > approach to take, so hoping that discussion lends some clarity to the > questions above. :) > > In the meantime I'll be working on the yellow items. > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Thu Jun 16 05:25:27 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Thu, 16 Jun 2016 01:25:27 -0400 Subject: [kontinuity-dev-public] The User Experience As It Stands In-Reply-To: References: Message-ID: On Wed, Jun 15, 2016 at 10:00 PM, Ben Parees wrote: > > > On Wed, Jun 15, 2016 at 5:07 PM, Andrew Lee Rubinger > wrote: > >> Hi all: >> >> OK, so at this point I think we have enough integration work done and >> user experience instructions set forth that I hope we're all starting to >> look at this from the same perspective. >> >> Apologies to any non-RHTers; I'm not sure how to share this outside the >> org at the moment: >> >> >> https://docs.google.com/a/redhat.com/document/d/1FtBt8XwZ237Mko6zEpkTOuFFhHfXM9THerjh702XA0c/edit?usp=sharing >> >> If everyone can center on that as the "thing to build out" and do dev >> against, that'd be wonderful. >> >> Orange are areas we want to improve but don't necessarily block us, and >> red is blocking us from saying "we're done here". >> >> Following those instructions, we get nicely down to observing the >> OpenShift project overview page for the application Catapult has just >> flung. Looks a little like this at the moment: >> >> https://imagebin.ca/v/2kniMSCFgVhS >> >> This is where I get a bit lost. :) >> >> We have 3 services: >> >> * Jenkins >> * jenkins-jnlp >> * kitchensink-html5-mobile (the example app in question) >> >> The example app then starts a build using s2i and *almost* completes: >> >> https://gist.github.com/ALRubinger/fd81272e2f8eaf0e07b8f4b61c8dab42 >> > > ?that's indicative of a problem w/ your cluster. i assume even if you > created a basic s2i buildconfig and kicked it off (no jenkins/pipelines > involved) you'd hit the same issue, so you'll need to dig into why your > registry appears to be inaccessible in your cluster. > > My Vagrantfile probably wasn't setting everything up correctly; Tomas' doesn't present this particular issue so I've updated the user instructions to switch to his. So now this builds correctly: https://gist.github.com/ALRubinger/68756e225a582b8a023a272fa33cee7f ...but there are no deployments associated so we just kinda dead-end. :) https://ibin.co/2kpCwGp25C1b.png > The Jenkins instance has 2 projects, "pipeline" and "pipeline-example", >> and I'm unsure how they relate to the example app we want to build/deploy. >> > > ?pipeline-example is built into the jenkins image, it's part of the basic > config we include in the image. we may get rid of it soon since it's > causing confusion, but it was intended as an out of the box "here's a > pipeline sample job": > > https://github.com/fabric8io/openshift-jenkins-s2i-config/tree/master/configuration/jobs/pipeline-example > > "pipeline" is, i assume, the name of your buildconfig that you're actually > interested in, that gets created for you when we see the buildconfig in > your openshift project. (I don't know for sure, the one we provide in the > tutorial is named "sample-pipeline", so i don't know where "pipeline" came > from in your case, but one way or another, you created it and it's the one > you care about) > > > > >> >> I have a sneaking suspicion that I've been informed that many of you are >> working on reconciling how this works together for your feature areas, but >> I'm not yet grokking the whole picture. :) >> >> 1) What's the intended role of the Jenkins service here? Shouldn't it be >> handling the pipeline build? >> > > ?the jenkins service provides access the the jenkins web console? and apis. > > > >> 2) Why is the build being kicked off in the example service itself? Is >> that what we want? >> > > ?not sure what you mean by a build being kicked off in the example > service.? The normal flow is: > 1) create an openshift buildconfig w/ strategy type pipeline > 2) a jenkins service+deployment gets created for you > 3) a new jenkins job gets created in the jenkins server that corresponds > to the buildconfig from (1) > 4) use openshift to start a build of the buildconfig from (1) > 5) that creates an openshift build object based on your buildconfig > 6) the fabric sync plugin running in the jenkins server from (2) detects > the new build object > 7) the fabric sync plugin running in the jenkins server starts the jenkins > job from (3) > Right, I think everything is happening correctly up to 6) or 7). When I say "kicked off in the example service" maybe my image link above will show --- the build is not happening in Jenkins. There is a "pipeline" job configured that does not run: https://ibin.co/2kpEYnwXP7Nr.png Manually triggering this job from the Jenkins UI results in indefinite blocking on an available executor: https://gist.github.com/ALRubinger/545af1e15aae13d4a23a09373316570b > 8) the jenkins pipeline job does whatever you defined it to do (which may > or may not include launching slave pods on openshift) > 9) the fabric sync plugin updates the openshift build object from (5) with > the ongoing state of the pipeline execution (what stage it's in, > success/failure, etc). > > > >> 3) Who's working on these pieces and are we all clear on what the >> experience should be? >> > > ?my team, jessica's team, the fabric team, your team. I feel pretty clear > about all the bits on the openshift/fabric/jenkins side. I haven't worried > about the Catapult layer since that's above those things as i understand it. > It's this integration I'd like everyone to jointly own and work on together until the "script" in the User Instructions completes. :) S, ALR > > >> >> I know Tomas and Xavier are syncing in the morning about the template >> file approach to take, so hoping that discussion lends some clarity to the >> questions above. :) >> >> In the meantime I'll be working on the yellow items. >> >> S, >> ALR >> >> -- >> Red Hat Developer Programs Architecture >> @ALRubinger >> >> _______________________________________________ >> kontinuity-dev-public mailing list >> kontinuity-dev-public at redhat.com >> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >> >> > > > -- > Ben Parees | OpenShift > > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Thu Jun 16 05:30:12 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Thu, 16 Jun 2016 01:30:12 -0400 Subject: [kontinuity-dev-public] OpenShift Template: Does it know the Host Binding? Message-ID: So I have a parameter in my Catapult template file: { "name": "CATAPULT_OPENSHIFT_CONSOLE_URL", "description": "The base URL of the OpenShift Console", "displayName": "OpenShift Console URL", "required": true, "value": "https://10.1.2.20:8443/console" } Is there a way I can default the value to something like https://${HOST_BIND_IP}/console ? Couldn't find a reference for built-in envvars available to the template. S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Thu Jun 16 06:22:52 2016 From: bparees at redhat.com (Ben Parees) Date: Wed, 15 Jun 2016 23:22:52 -0700 Subject: [kontinuity-dev-public] OpenShift Template: Does it know the Host Binding? In-Reply-To: References: Message-ID: On Jun 15, 2016 22:30, "Andrew Lee Rubinger" wrote: > > So I have a parameter in my Catapult template file: > > { > "name": "CATAPULT_OPENSHIFT_CONSOLE_URL", > "description": "The base URL of the OpenShift Console", > "displayName": "OpenShift Console URL", > "required": true, > "value": "https://10.1.2.20:8443/console" > } > > Is there a way I can default the value to something like https://${HOST_BIND_IP}/console ? Couldn't find a reference for built-in envvars available to the template. Unfortunately not. > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Thu Jun 16 06:47:52 2016 From: bparees at redhat.com (Ben Parees) Date: Wed, 15 Jun 2016 23:47:52 -0700 Subject: [kontinuity-dev-public] The User Experience As It Stands In-Reply-To: References: Message-ID: On Wed, Jun 15, 2016 at 10:25 PM, Andrew Lee Rubinger wrote: > > > On Wed, Jun 15, 2016 at 10:00 PM, Ben Parees wrote: > >> >> >> On Wed, Jun 15, 2016 at 5:07 PM, Andrew Lee Rubinger >> wrote: >> >>> Hi all: >>> >>> OK, so at this point I think we have enough integration work done and >>> user experience instructions set forth that I hope we're all starting to >>> look at this from the same perspective. >>> >>> Apologies to any non-RHTers; I'm not sure how to share this outside the >>> org at the moment: >>> >>> >>> https://docs.google.com/a/redhat.com/document/d/1FtBt8XwZ237Mko6zEpkTOuFFhHfXM9THerjh702XA0c/edit?usp=sharing >>> >>> If everyone can center on that as the "thing to build out" and do dev >>> against, that'd be wonderful. >>> >>> Orange are areas we want to improve but don't necessarily block us, and >>> red is blocking us from saying "we're done here". >>> >>> Following those instructions, we get nicely down to observing the >>> OpenShift project overview page for the application Catapult has just >>> flung. Looks a little like this at the moment: >>> >>> https://imagebin.ca/v/2kniMSCFgVhS >>> >>> This is where I get a bit lost. :) >>> >>> We have 3 services: >>> >>> * Jenkins >>> * jenkins-jnlp >>> * kitchensink-html5-mobile (the example app in question) >>> >>> The example app then starts a build using s2i and *almost* completes: >>> >>> https://gist.github.com/ALRubinger/fd81272e2f8eaf0e07b8f4b61c8dab42 >>> >> >> ?that's indicative of a problem w/ your cluster. i assume even if you >> created a basic s2i buildconfig and kicked it off (no jenkins/pipelines >> involved) you'd hit the same issue, so you'll need to dig into why your >> registry appears to be inaccessible in your cluster. >> >> > My Vagrantfile probably wasn't setting everything up correctly; Tomas' > doesn't present this particular issue so I've updated the user instructions > to switch to his. > > So now this builds correctly: > > https://gist.github.com/ALRubinger/68756e225a582b8a023a272fa33cee7f > > ...but there are no deployments associated so we just kinda dead-end. :) > > https://ibin.co/2kpCwGp25C1b.png > ?ok..do you need help constructing a DC that will be triggered by the new image produced from the build?? > > >> The Jenkins instance has 2 projects, "pipeline" and "pipeline-example", >>> and I'm unsure how they relate to the example app we want to build/deploy. >>> >> >> ?pipeline-example is built into the jenkins image, it's part of the basic >> config we include in the image. we may get rid of it soon since it's >> causing confusion, but it was intended as an out of the box "here's a >> pipeline sample job": >> >> https://github.com/fabric8io/openshift-jenkins-s2i-config/tree/master/configuration/jobs/pipeline-example >> >> "pipeline" is, i assume, the name of your buildconfig that you're >> actually interested in, that gets created for you when we see the >> buildconfig in your openshift project. (I don't know for sure, the one we >> provide in the tutorial is named "sample-pipeline", so i don't know where >> "pipeline" came from in your case, but one way or another, you created it >> and it's the one you care about) >> >> >> >> >>> >>> I have a sneaking suspicion that I've been informed that many of you are >>> working on reconciling how this works together for your feature areas, but >>> I'm not yet grokking the whole picture. :) >>> >>> 1) What's the intended role of the Jenkins service here? Shouldn't it >>> be handling the pipeline build? >>> >> >> ?the jenkins service provides access the the jenkins web console? and >> apis. >> >> >> >>> 2) Why is the build being kicked off in the example service itself? Is >>> that what we want? >>> >> >> ?not sure what you mean by a build being kicked off in the example >> service.? The normal flow is: >> 1) create an openshift buildconfig w/ strategy type pipeline >> 2) a jenkins service+deployment gets created for you >> 3) a new jenkins job gets created in the jenkins server that corresponds >> to the buildconfig from (1) >> 4) use openshift to start a build of the buildconfig from (1) >> 5) that creates an openshift build object based on your buildconfig >> 6) the fabric sync plugin running in the jenkins server from (2) detects >> the new build object >> 7) the fabric sync plugin running in the jenkins server starts the >> jenkins job from (3) >> > > Right, I think everything is happening correctly up to 6) or 7). When I > say "kicked off in the example service" maybe my image link above will show > --- the build is not happening in Jenkins. There is a "pipeline" job > configured that does not run: > > https://ibin.co/2kpEYnwXP7Nr.png > > Manually triggering this job from the Jenkins UI results in indefinite > blocking on an available executor: > > https://gist.github.com/ALRubinger/545af1e15aae13d4a23a09373316570b > ?Take a look at the pod logs for the jenkins pod, they will give you some clue what is failing, in the form of exceptions. > > >> 8) the jenkins pipeline job does whatever you defined it to do (which may >> or may not include launching slave pods on openshift) >> 9) the fabric sync plugin updates the openshift build object from (5) >> with the ongoing state of the pipeline execution (what stage it's in, >> success/failure, etc). >> >> >> >>> 3) Who's working on these pieces and are we all clear on what the >>> experience should be? >>> >> >> ?my team, jessica's team, the fabric team, your team. I feel pretty >> clear about all the bits on the openshift/fabric/jenkins side. I haven't >> worried about the Catapult layer since that's above those things as i >> understand it. >> > > It's this integration I'd like everyone to jointly own and work on > together until the "script" in the User Instructions completes. :) > > S, > ALR > > >> >> >>> >>> I know Tomas and Xavier are syncing in the morning about the template >>> file approach to take, so hoping that discussion lends some clarity to the >>> questions above. :) >>> >>> In the meantime I'll be working on the yellow items. >>> >>> S, >>> ALR >>> >>> -- >>> Red Hat Developer Programs Architecture >>> @ALRubinger >>> >>> _______________________________________________ >>> kontinuity-dev-public mailing list >>> kontinuity-dev-public at redhat.com >>> https://www.redhat.com/mailman/listinfo/kontinuity-dev-public >>> >>> >> >> >> -- >> Ben Parees | OpenShift >> >> > > > -- > Red Hat Developer Programs Architecture > @ALRubinger > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Fri Jun 17 14:27:18 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Fri, 17 Jun 2016 16:27:18 +0200 Subject: [kontinuity-dev-public] Jenkins service account does not work on slaves (only on master) Message-ID: <1466173638.3478.54.camel@redhat.com> I have modified my Jenkinsfile to use service account and I have found out that Jenkins master and slaves run under different service accounts.? Ben's template creates ServiceAccount named "jenkins" with the correct permissions but slaves are run with ServiceAccount named "default" without permissions. This ends up for me after login with: $ oc login https://kubernetes.default/ --token=$(cat /run/secrets/kubernetes.io/serviceaccount/token) --certificate- authority=/run/secrets/kubernetes.io/serviceaccount/ca.crt ? "Logged into "https://kubernetes.default:443" as system:serviceaccount:catapult:default using the token provided. ? "You don't have any projects. ..." Changing to the same project I am in ends with: $ oc project catapult error: You are not a member of project "catapult". You are not a member of any projects. You can request a project to be created with the 'new-project' command. I tried those steps manually by "docker exec" into master instance and it works fine there with "jenkins" service account. Can I somehow set up Jenkins slaves to use the correct service account? Or can we roll back to not having special service account for Jenkins and use the default one? I think it was in the template before... Thanks, Tomas -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Sun Jun 19 02:05:56 2016 From: bparees at redhat.com (Ben Parees) Date: Sat, 18 Jun 2016 19:05:56 -0700 Subject: [kontinuity-dev-public] Jenkins service account does not work on slaves (only on master) In-Reply-To: <1466173638.3478.54.camel@redhat.com> References: <1466173638.3478.54.camel@redhat.com> Message-ID: On Fri, Jun 17, 2016 at 7:27 AM, Tomas Nozicka wrote: > I have modified my Jenkinsfile to use service account and I have found out > that Jenkins master and slaves run under different service accounts. > > Ben's template creates ServiceAccount named "jenkins" with the correct > permissions but slaves are run with ServiceAccount named "default" without > permissions. This ends up for me after login with: > $ oc login https://kubernetes.default/ --token=$(cat /run/secrets/ > kubernetes.io/serviceaccount/token) --certificate-authority=/run/secrets/ > kubernetes.io/serviceaccount/ca.crt > "Logged into "https://kubernetes.default:443" as > system:serviceaccount:catapult:default using the token provided. > "You don't have any projects. ..." > > Changing to the same project I am in ends with: > $ oc project catapult > error: You are not a member of project "catapult". > You are not a member of any projects. You can request a project to be > created with the 'new-project' command. > > I tried those steps manually by "docker exec" into master instance and it > works fine there with "jenkins" service account. > > Can I somehow set up Jenkins slaves to use the correct service account? > Or can we roll back to not having special service account for Jenkins and > use the default one? I think it was in the template before... > ?using the default service account means we have to grant edit permission to the default service account, which is undesirable. Sounds like we need a patch to the k8s plugin to use the same service account the master is using and/or make the service account configurable as part of the podtemplate definition. Jimmi, what does the kubernetes-workflow plugin do with respect to this situation? ? ?in the meantime, if you grant edit permissions to your default svc account, you should be set.? > > Thanks, > Tomas > -- Ben Parees | OpenShift -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Sun Jun 19 11:05:42 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Sun, 19 Jun 2016 13:05:42 +0200 Subject: [kontinuity-dev-public] Jenkins service account does not work on slaves (only on master) In-Reply-To: References: <1466173638.3478.54.camel@redhat.com> Message-ID: <1466334342.6065.6.camel@redhat.com> On So, 2016-06-18 at 19:05 -0700, Ben Parees wrote: > > > On Fri, Jun 17, 2016 at 7:27 AM, Tomas Nozicka > wrote: > > I have modified my Jenkinsfile to use service account and I have > > found out that Jenkins master and slaves run under different > > service accounts.? > > > > Ben's template creates ServiceAccount named "jenkins" with the > > correct permissions but slaves are run with ServiceAccount named > > "default" without permissions. This ends up for me after login > > with: > > $ oc login https://kubernetes.default/ --token=$(cat > > /run/secrets/kubernetes.io/serviceaccount/token) --certificate- > > authority=/run/secrets/kubernetes.io/serviceaccount/ca.crt > > ? "Logged into "https://kubernetes.default:443" as > > system:serviceaccount:catapult:default using the token provided. > > ? "You don't have any projects. ..." > > > > Changing to the same project I am in ends with: > > $ oc project catapult > > error: You are not a member of project "catapult". > > You are not a member of any projects. You can request a project to > > be created with the 'new-project' command. > > > > I tried those steps manually by "docker exec" into master instance > > and it works fine there with "jenkins" service account. > > > > Can I somehow set up Jenkins slaves to use the correct service > > account? > > Or can we roll back to not having special service account for > > Jenkins and use the default one? I think it was in the template > > before... > > > using the default service account means we have to grant edit > permission to the default service account, which is undesirable.? > Sounds like we need a patch to the k8s plugin to use the same service > account the master is using and/or make the service account > configurable as part of the podtemplate definition. > > Jimmi, what does the kubernetes-workflow plugin do with respect to > this situation? > > in the meantime, if you grant edit permissions to your default svc > account, you should be set. Thanks Ben, I have actually went with other workaround: #!/usr/bin/groovy def k8s_token = '' node ('master') { ????k8s_token = readFile '/var/run/secrets/kubernetes.io/serviceaccount/token' } node('redhatdistortion-maven') { ? ? sh "oc login https://kubernetes.default/ --token=${k8s_token} -- certificate-authority=/run/secrets/kubernetes.io/serviceaccount/ca.crt" } But it is still a workaround. Do you think we could think add ?some permission to jenkins service account to allow this: + oc policy add-role-to-user system:image-puller system:serviceaccount:test-prod:default --namespace=test Error from server: User "system:serviceaccount:test:jenkins" cannot get policybindings in project "test" It is the last think that does not work with using service accounts right now in my pipeline. Thanks, Tomas > > ? > > > > Thanks, > > Tomas > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Sun Jun 19 16:31:50 2016 From: bparees at redhat.com (Ben Parees) Date: Sun, 19 Jun 2016 09:31:50 -0700 Subject: [kontinuity-dev-public] Jenkins service account does not work on slaves (only on master) In-Reply-To: <1466334342.6065.6.camel@redhat.com> References: <1466173638.3478.54.camel@redhat.com> <1466334342.6065.6.camel@redhat.com> Message-ID: On Jun 19, 2016 4:05 AM, "Tomas Nozicka" wrote: > > On So, 2016-06-18 at 19:05 -0700, Ben Parees wrote: >> >> >> >> On Fri, Jun 17, 2016 at 7:27 AM, Tomas Nozicka wrote: >>> >>> I have modified my Jenkinsfile to use service account and I have found out that Jenkins master and slaves run under different service accounts. >>> >>> Ben's template creates ServiceAccount named "jenkins" with the correct permissions but slaves are run with ServiceAccount named "default" without permissions. This ends up for me after login with: >>> $ oc login https://kubernetes.default/ --token=$(cat /run/secrets/ kubernetes.io/serviceaccount/token) --certificate-authority=/run/secrets/ kubernetes.io/serviceaccount/ca.crt >>> "Logged into "https://kubernetes.default:443" as system:serviceaccount:catapult:default using the token provided. >>> "You don't have any projects. ..." >>> >>> Changing to the same project I am in ends with: >>> $ oc project catapult >>> error: You are not a member of project "catapult". >>> You are not a member of any projects. You can request a project to be created with the 'new-project' command. >>> >>> I tried those steps manually by "docker exec" into master instance and it works fine there with "jenkins" service account. >>> >>> Can I somehow set up Jenkins slaves to use the correct service account? >>> Or can we roll back to not having special service account for Jenkins and use the default one? I think it was in the template before... >>> >> using the default service account means we have to grant edit permission to the default service account, which is undesirable. Sounds like we need a patch to the k8s plugin to use the same service account the master is using and/or make the service account configurable as part of the podtemplate definition. >> >> Jimmi, what does the kubernetes-workflow plugin do with respect to this situation? >> >> in the meantime, if you grant edit permissions to your default svc account, you should be set. > > Thanks Ben, I have actually went with other workaround: > > #!/usr/bin/groovy > def k8s_token = '' > node ('master') { > k8s_token = readFile '/var/run/secrets/ kubernetes.io/serviceaccount/token' > } > node('redhatdistortion-maven') { > sh "oc login https://kubernetes.default/ --token=${k8s_token} --certificate-authority=/run/secrets/kubernetes.io/serviceaccount/ca.crt" > } > > But it is still a workaround. > > Do you think we could think add some permission to jenkins service account to allow this: > > + oc policy add-role-to-user system:image-puller system:serviceaccount:test-prod:default --namespace=test No I don't think we can do that, that's a specific grant to a specific project that not all users will have. We can only add permissions relative to the current project(in your case, catapult I believe). There are really two problems: 1) the user may not have permission to grant themselves access to another project (you are presumably running that command as cluster admin) and 2) the project name is specific to your scenario. So this is something you'll have to set up explicitly. > > Error from server: User "system:serviceaccount:test:jenkins" cannot get policybindings in project "test" > > > It is the last think that does not work with using service accounts right now in my pipeline. > > Thanks, > Tomas > >> >> >>> >>> >>> Thanks, >>> Tomas >>> >> >> -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Sun Jun 19 18:18:27 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Sun, 19 Jun 2016 20:18:27 +0200 Subject: [kontinuity-dev-public] Jenkins service account does not work on slaves (only on master) In-Reply-To: References: <1466173638.3478.54.camel@redhat.com> <1466334342.6065.6.camel@redhat.com> Message-ID: <1466360307.6065.25.camel@redhat.com> Inline. On Ne, 2016-06-19 at 09:31 -0700, Ben Parees wrote: > > Do you think we could add some permission to jenkins service > account to allow this: > > > > + oc policy add-role-to-user system:image-puller > system:serviceaccount:test-prod:default --namespace=test > No I don't think we can do that, that's a specific grant to a > specific project that not all users will have. We can only add > permissions relative to the current project(in your case, catapult I > believe). There are really two problems: 1) the user may not have > permission to grant themselves access to another project (you are > presumably running that command as cluster admin) and 2) the project > name is specific to your scenario. > So this is something you'll have to set up explicitly. I feel like you misunderstood?my point and I should have been clearer about it.? I don't want you to setup that access to another project. I am perfectly ok with issuing that command by myself: ?oc policy add-role-to-user system:image- puller?system:serviceaccount:test-prod:default --namespace=test What I have asked you to do is to setup RoleBinding in jenkins template, like: ??https://github.com/openshift/origin/blob/master/examples/jenkins/pipe line/jenkinstemplate.json#L153 You currently set up only RoleBinding to role "edit". And that does not seem to be enough, because I get error: ??Error from server: User "system:serviceaccount:test:jenkins" cannot?get policybindings in project "test" And I am not sure which one is needed for this type of action. And no, I don't use admin account there. I have created first project by UI logged as user "openshift-dev" and the second namespace was created by service account jenkins from the first namespace. I can create any services, deployments, ... in both projects through the service account so I already have permissions to do almost anything to those projects. But cannot connect one to each other so the deployment in second namespace can be allowed to deploy from ImageStream in first namespace. > > > > Error from server: User "system:serviceaccount:test:jenkins" cannot > get policybindings in project "test" > > > > > > It is the last think that does not work with using service accounts > right now in my pipeline. > > > > Thanks, > > Tomas > > > >> > >> ? > >>> > >>> > >>> Thanks, > >>> Tomas > >>> > >> > >> From tnozicka at redhat.com Mon Jun 20 08:52:48 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Mon, 20 Jun 2016 10:52:48 +0200 Subject: [kontinuity-dev-public] Jenkins service account does not work on slaves (only on master) In-Reply-To: <1466360307.6065.25.camel@redhat.com> References: <1466173638.3478.54.camel@redhat.com> <1466334342.6065.6.camel@redhat.com> <1466360307.6065.25.camel@redhat.com> Message-ID: <1466412768.4033.13.camel@redhat.com> On Ne, 2016-06-19 at 20:18 +0200, Tomas Nozicka wrote: > Inline. > > On Ne, 2016-06-19 at 09:31 -0700, Ben Parees wrote: > > > > > > > > Do you think we could add some permission to jenkins service > > account to allow this: > > > > > > > > > + oc policy add-role-to-user system:image-puller > > system:serviceaccount:test-prod:default --namespace=test > > No I don't think we can do that, that's a specific grant to a > > specific project that not all users will have. We can only add > > permissions relative to the current project(in your case, catapult > > I > > believe). There are really two problems: 1) the user may not have > > permission to grant themselves access to another project (you are > > presumably running that command as cluster admin) and 2) the > > project > > name is specific to your scenario. > > So this is something you'll have to set up explicitly. > I feel like you misunderstood?my point and I should have been clearer > about it.? > > I don't want you to setup that access to another project. I am > perfectly ok with issuing that command by myself: > > ?oc policy add-role-to-user system:image- > puller?system:serviceaccount:test-prod:default --namespace=test > > What I have asked you to do is to setup RoleBinding in jenkins > template, like: > ??https://github.com/openshift/origin/blob/master/examples/jenkins/pi > pe > line/jenkinstemplate.json#L153 > You currently set up only RoleBinding to role "edit". And that does > not > seem to be enough, because I get error: > > ??Error from server: User "system:serviceaccount:test:jenkins" > cannot?get policybindings in project "test" > > And I am not sure which one is needed for this type of action. > And no, I don't use admin account there. I have created first project > by UI logged as user "openshift-dev" and the second namespace was > created by service account jenkins from the first namespace. > I can create any services, deployments, ... in both projects through > the service account so I already have permissions to do almost > anything > to those projects. But cannot connect one to each other so the > deployment in second namespace can be allowed to deploy from > ImageStream in first namespace. The last paragraph I wrote is not true. I must have had modified cluster policies or did something else :( I have recreated my cluster and I am running into (justified) permission denials. I am really sorry about the confusion. So considering this is the week, I am not going to pursue the idea of separating dev from prod through namespace for now. Thanks everybody for their help so far! -- Tomas > > > > > > > > > > > > > Error from server: User "system:serviceaccount:test:jenkins" > > > cannot > > get policybindings in project "test" > > > > > > > > > > > > It is the last think that does not work with using service > > > accounts > > right now in my pipeline. > > > > > > > > > Thanks, > > > Tomas > > > > > > > > > > > > > > > ? > > > > > > > > > > > > > > > > > > > > Thanks, > > > > > Tomas > > > > > > > > > From alr at redhat.com Mon Jun 20 14:36:01 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Mon, 20 Jun 2016 10:36:01 -0400 Subject: [kontinuity-dev-public] Jenkins service account does not work on slaves (only on master) In-Reply-To: <1466412768.4033.13.camel@redhat.com> References: <1466173638.3478.54.camel@redhat.com> <1466334342.6065.6.camel@redhat.com> <1466360307.6065.25.camel@redhat.com> <1466412768.4033.13.camel@redhat.com> Message-ID: On Jun 20, 2016 4:52 AM, "Tomas Nozicka" wrote: > > On Ne, 2016-06-19 at 20:18 +0200, Tomas Nozicka wrote: > > Inline. > > > > On Ne, 2016-06-19 at 09:31 -0700, Ben Parees wrote: > > > > > > > > > > > Do you think we could add some permission to jenkins service > > > account to allow this: > > > > > > > > > > > > + oc policy add-role-to-user system:image-puller > > > system:serviceaccount:test-prod:default --namespace=test > > > No I don't think we can do that, that's a specific grant to a > > > specific project that not all users will have. We can only add > > > permissions relative to the current project(in your case, catapult > > > I > > > believe). There are really two problems: 1) the user may not have > > > permission to grant themselves access to another project (you are > > > presumably running that command as cluster admin) and 2) the > > > project > > > name is specific to your scenario. > > > So this is something you'll have to set up explicitly. > > I feel like you misunderstood my point and I should have been clearer > > about it. > > > > I don't want you to setup that access to another project. I am > > perfectly ok with issuing that command by myself: > > > > oc policy add-role-to-user system:image- > > puller system:serviceaccount:test-prod:default --namespace=test > > > > What I have asked you to do is to setup RoleBinding in jenkins > > template, like: > > https://github.com/openshift/origin/blob/master/examples/jenkins/pi > > pe > > line/jenkinstemplate.json#L153 > > You currently set up only RoleBinding to role "edit". And that does > > not > > seem to be enough, because I get error: > > > > Error from server: User "system:serviceaccount:test:jenkins" > > cannot get policybindings in project "test" > > > > And I am not sure which one is needed for this type of action. > > > > > And no, I don't use admin account there. I have created first project > > by UI logged as user "openshift-dev" and the second namespace was > > created by service account jenkins from the first namespace. > > I can create any services, deployments, ... in both projects through > > the service account so I already have permissions to do almost > > anything > > to those projects. But cannot connect one to each other so the > > deployment in second namespace can be allowed to deploy from > > ImageStream in first namespace. > The last paragraph I wrote is not true. I must have had modified > cluster policies or did something else :( I have recreated my cluster > and I am running into (justified) permission denials. I am really sorry > about the confusion. > > So considering this is the week, I am not going to pursue the idea of > separating dev from prod through namespace for now. +1 to both pursuing the separation ultimately, and cutting it from the release criteria for Summit. Creating projects on behalf of users is something we'll have to get in for OpenShift Online. Thanks for starting work on it! > > Thanks everybody for their help so far! > -- > Tomas > > > > > > > > > > > > > > > > > > > > Error from server: User "system:serviceaccount:test:jenkins" > > > > cannot > > > get policybindings in project "test" > > > > > > > > > > > > > > > > It is the last think that does not work with using service > > > > accounts > > > right now in my pipeline. > > > > > > > > > > > > Thanks, > > > > Tomas > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Thanks, > > > > > > Tomas > > > > > > > > > > > > > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public -------------- next part -------------- An HTML attachment was scrubbed... URL: From ashley.benson at edatadeals.com Mon Jun 20 14:45:48 2016 From: ashley.benson at edatadeals.com (Ashley Benson) Date: Mon, 20 Jun 2016 20:15:48 +0530 Subject: [kontinuity-dev-public] Amazon Global Cloud Users Contacts Message-ID: Hello, My Name is Ashley Benson and I am Marketing Manager. I have gone through your website. Would be interested in acquiring "Amazon Global Cloud Users" List for your Email Marketing, Email Campaign and we have Other Technology Users List with emails and complete contact information. Cloud Users: Google, Soft Layer, IBM Marketing Cloud, Bluemix, Hybrid, IaaS -- Infrastructure as a Service, PaaS -- Platform as a Service, SaaS -- Software as a Service and many more. We have list of other Technologies users: MSoft Users, Salesforce Users, SAP Users, Citrix users, XBLR Users, OneStream Users, ECM, ERM, Solidwork Users, VMware Users, ERP/CRM/PLM Users and many more. Do share your target audience like Technology/ Industry / Geography / Job title, So that we can give you more information about our services. Our List Includes:- Company Name, Web Address, Contact Name, Verified Email, Job Title, Application Type, Complete Mailing Address, FAX Number, Total Employees, SIC Code and Industry details. Kindly let me know if you would be interested in purchasing this information and I shall provide you further details. We will be happy to provide you with some of the samples for your validation. Thanks and looking forward to hear from you! Best Regards, Ashley Benson Marketing Manager If you do not wish to receive further emails kindly reply with Unsubscribe -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Mon Jun 20 15:51:14 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Mon, 20 Jun 2016 11:51:14 -0400 Subject: [kontinuity-dev-public] Exposing OpenShift from ADB Publicly Message-ID: Hi, all: So I have a Fedora23 server at rhtsummitdevnation.alrubinger.com configured with libvirt/Vagrant and running our OpenShift on ADB setup: https://github.com/redhat-kontinuity/adb-openshift-runtime/blob/master/Vagrantfile This is binding to an internal IP: 10.1.2.2 Now, we can route requests into the machine's external IP to map to 10.1.2.2, but OpenShift (and the rest of the application layer) is configured to use the internal IP and will reference out to the user routes like *.10.2.2.xip.io, for example. Also the redirects in the console... Andrew Block notes that he's got the juju to configure OpenShift and ADB to make things available externally, so I'll let him take it from here for what we need to do? RHT employees: contact me for root access to the server if you'd like to help. S ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From monica at discoverlists.com Mon Jun 20 17:48:28 2016 From: monica at discoverlists.com (Monica Aguilera) Date: Mon, 20 Jun 2016 23:18:28 +0530 Subject: [kontinuity-dev-public] Oracle Applications end users Message-ID: <823ab640-fd84-a88b-563c-73192894740c@discoverlists.com> -- Hi, Would you be interested in reaching out to our recently verified *IT and Software Industry Professionals**using various Oracle Applications*? _*Job titles Include*___:- Executive Management (C-level, President, EVP, SVP), Architect, Partner (Member of Oracle Partner Network), Database Administrator, Applications Manager, Operations Manager, LOB Manager/Director, Data Warehouse Manager, Developer, Consultant, Business Analyst etc. You can get an immediate access to this targeted database for your Telemarketing, Email Marketing or any other Direct marketing Campaign. *About us:*We are one of the influential leading global database solution providers, offering B2B permission based contacts as per your requirement. Awaiting your valuable response. Note: List can be customized based on your Target Requirements. *Regards,* Monica Aguilera, Marketing Associate. /If you don't wish to receive any offer from us please feel free to reply to this email with "Leave Out" / -------------- next part -------------- An HTML attachment was scrubbed... URL: From sophie.wintrich at gdatalist.com Mon Jun 20 18:07:27 2016 From: sophie.wintrich at gdatalist.com (Sophie wintrich) Date: Mon, 20 Jun 2016 13:07:27 -0500 Subject: [kontinuity-dev-public] Redhat users contacts Lists 2016 Message-ID: <0ffa01d1cb1e$9b0ca690$d125f3b0$@gdatalist.com> Hi, Would you be interested in Acquiring Redhat Users contacts list of 2016? Information Fields - First and Last name, Phone number, Email Address, Company Name, Job Title, Address, City, State, Zip, SIC code/Industry, Revenue and Company Size. The leads can also be further customized as per requirements. We can provide contact list from any country/industry/title. We also have the following Users: * Ubnntu * Debian * Suse * Centos Please review and let me know if you are interested in any of the technology users or different contact list for your campaigns and I will provide more information for the same. Appreciate your time and look forward to hear from you. Thanks, Sophie wintrich Demand Generation-Technology Database | List acquisition | Technology Lists | Email/Data Appending | Search Engine Optimization | To Opt Out, please reply with Leave Out in the Subject Line. -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Tue Jun 21 03:42:03 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Mon, 20 Jun 2016 23:42:03 -0400 Subject: [kontinuity-dev-public] Exposing OpenShift from ADB Publicly In-Reply-To: References: Message-ID: Much closer today thanks to the work done by Andrew and Edson! Though not seeming to be able to connect now: https://206.51.239.218:8443/ ...and service firewalld status -l is showing some interesting stuff: Jun 20 23:39:56 rhtsummitdevnation.alrubinger.com /firewalld[28916]: ERROR: COMMAND_FAILED: '/usr/sbin/ip6tables -w --table filter --delete FORWARD --in-interface virbr2 --jump REJECT' failed: ip6tables: No chain/target/match by that name Let's pick this up tomorrow? S, ALR On Mon, Jun 20, 2016 at 11:51 AM, Andrew Lee Rubinger wrote: > Hi, all: > > So I have a Fedora23 server at rhtsummitdevnation.alrubinger.com > configured with libvirt/Vagrant and running our OpenShift on ADB setup: > > > https://github.com/redhat-kontinuity/adb-openshift-runtime/blob/master/Vagrantfile > > This is binding to an internal IP: 10.1.2.2 > > Now, we can route requests into the machine's external IP to map to > 10.1.2.2, but OpenShift (and the rest of the application layer) is > configured to use the internal IP and will reference out to the user routes > like *.10.2.2.xip.io, for example. Also the redirects in the console... > > Andrew Block notes that he's got the juju to configure OpenShift and ADB > to make things available externally, so I'll let him take it from here for > what we need to do? > > RHT employees: contact me for root access to the server if you'd like to > help. > > S > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Tue Jun 21 05:51:41 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Tue, 21 Jun 2016 01:51:41 -0400 Subject: [kontinuity-dev-public] Exposing OpenShift from ADB Publicly In-Reply-To: References: Message-ID: Ah I see, I'd re-pulled the new version of the Vagrantfile and Andrew has left instructions for how the Vagrantfile needs changes to enable ADB/OpenShift to be brought to the light of the outside world. Will look to apply those tomorrow and report back. S, ALR On Mon, Jun 20, 2016 at 11:42 PM, Andrew Lee Rubinger wrote: > Much closer today thanks to the work done by Andrew and Edson! > > Though not seeming to be able to connect now: > > https://206.51.239.218:8443/ > > ...and service firewalld status -l is showing some interesting stuff: > > Jun 20 23:39:56 rhtsummitdevnation.alrubinger.com /firewalld[28916]: > ERROR: COMMAND_FAILED: '/usr/sbin/ip6tables -w --table filter --delete > FORWARD --in-interface virbr2 --jump REJECT' failed: ip6tables: No > chain/target/match by that name > > Let's pick this up tomorrow? > > S, > ALR > > On Mon, Jun 20, 2016 at 11:51 AM, Andrew Lee Rubinger > wrote: > >> Hi, all: >> >> So I have a Fedora23 server at rhtsummitdevnation.alrubinger.com >> configured with libvirt/Vagrant and running our OpenShift on ADB setup: >> >> >> https://github.com/redhat-kontinuity/adb-openshift-runtime/blob/master/Vagrantfile >> >> This is binding to an internal IP: 10.1.2.2 >> >> Now, we can route requests into the machine's external IP to map to >> 10.1.2.2, but OpenShift (and the rest of the application layer) is >> configured to use the internal IP and will reference out to the user routes >> like *.10.2.2.xip.io, for example. Also the redirects in the console... >> >> Andrew Block notes that he's got the juju to configure OpenShift and ADB >> to make things available externally, so I'll let him take it from here for >> what we need to do? >> >> RHT employees: contact me for root access to the server if you'd like to >> help. >> >> S >> ALR >> >> -- >> Red Hat Developer Programs Architecture >> @ALRubinger >> > > > > -- > Red Hat Developer Programs Architecture > @ALRubinger > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Thu Jun 23 10:29:11 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Thu, 23 Jun 2016 06:29:11 -0400 Subject: [kontinuity-dev-public] Today's IRC Scrum Message-ID: Hi, all: We're almost there! I'll likely not be at the IRC Scrum today, but everyone can feel free to use the calendar block to catch up on open issues. My test runs tonight are showing me, with a couple small things awaiting fixes, a viable product for Summit. :D Open stuff: https://docs.google.com/a/redhat.com/document/d/1FtBt8XwZ237Mko6zEpkTOuFFhHfXM9THerjh702XA0c/edit?usp=sharing (RHT only, sorry!) * Clayton: Can you look into getting us another Origin release 1.3.0-alpha.3? * Ben ** https://github.com/openshift/jenkins/pull/116 * Tomas ** Review Rupali's NodeJS example ** kitchensink-html5-mobile pipelined up ** GitHub webhooks build trigger * Me ** Figure out some presentation to make everyone's work look as good as it is. I *believe* the Fabric8 team, UXD, and the Origin Console team have met acceptance criteria for their deliverables for the Theremin release! Pending anything we discover that needs patching/fixing. Later on for RHT-ers I'll record the user flow to show everyone what the integrated experience looks like so far. :) S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Sat Jun 25 02:30:19 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Fri, 24 Jun 2016 22:30:19 -0400 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely Message-ID: Hi all: So with Ben's new changes in place (thanks!) the k8s namespace is set in Jenkins kubernetes-plugin: https://ibin.co/2lpVENzMzOGm.png This gets us an OpenShift build triggered properly, but it waits indefinitely without ever kicking off the build in Jenkins: https://ibin.co/2lpVJNi1p37O.png Manually triggering a second build in OpenShift calls into Jenkins properly and completes as expected. So...race condition? Tries to kick off a build before namespace is set? S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Sat Jun 25 02:57:20 2016 From: bparees at redhat.com (Ben Parees) Date: Fri, 24 Jun 2016 22:57:20 -0400 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely In-Reply-To: References: Message-ID: On Jun 24, 2016 10:30 PM, "Andrew Lee Rubinger" wrote: > > Hi all: > > So with Ben's new changes in place (thanks!) the k8s namespace is set in Jenkins kubernetes-plugin: > > https://ibin.co/2lpVENzMzOGm.png > > This gets us an OpenShift build triggered properly, but it waits indefinitely without ever kicking off the build in Jenkins: > > https://ibin.co/2lpVJNi1p37O.png > > Manually triggering a second build in OpenShift calls into Jenkins properly and completes as expected. > > So...race condition? Tries to kick off a build before namespace is set? No, the namespace is set in the jenkins config.xml before jenkins after starts, so it's not that. (also it only affects slave launching, not pipeline starting) But what it is i'm not sure. The first build is created/started by you instantiating your template? I think it's a question for jimmi as to why it doesn't see the new openshift build and start the pipeline in response. > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Sat Jun 25 05:28:21 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Sat, 25 Jun 2016 01:28:21 -0400 Subject: [kontinuity-dev-public] Webhooks to our remote OpenShift instance Message-ID: So we've equipped Catapult with the ability to add a webhook to the GitHub repo so it kicks OpenShift when there's a push (yay!). It looks a little like: https://147.75.202.53:8443/oapi/v1/namespaces/jboss-eap-quickstarts/buildconfigs/helloworld-pipeline/webhooks/kontinu8/github Problem is, GitHub isn't liking it. :) "We had a problem connecting to the server. The most common problem with these types of errors is a misconfigured SSL certificate. Try verifying your certificate chain." Short of installing a proper certificate on this server and giving it a domain, etc; any bright ideas to clear this up so we can get autobuild on push events? S, ALR -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From ablock at redhat.com Sat Jun 25 05:36:13 2016 From: ablock at redhat.com (Andrew Block) Date: Sat, 25 Jun 2016 00:36:13 -0500 Subject: [kontinuity-dev-public] Webhooks to our remote OpenShift instance In-Reply-To: References: Message-ID: The GitHub API [1] supports setting an insecure flag to disable SSL validation in web hooks. The java client does allow you to configure this setting in the config field [2]. Whether we should allow that is another story, but it is an option. - Andy [1] -?https://developer.github.com/v3/repos/hooks/? [2] -?https://github.com/redhat-kontinuity/catapult/blob/master/services/github-service-impl/src/main/java/org/kontinuity/catapult/service/github/impl/kohsuke/KohsukeGitHubServiceImpl.java#L184-L186? ? Andrew Block Red Hat Consulting 101 N. Wacker, Suite 150 Chicago, IL 60606 andrew.block at redhat.com?| m. (716) 870-2408 On June 25, 2016 at 12:28:54 AM, Andrew Lee Rubinger (alr at redhat.com) wrote: So we've equipped Catapult with the ability to add a webhook to the GitHub repo so it kicks OpenShift when there's a push (yay!). It looks a little like: ? https://147.75.202.53:8443/oapi/v1/namespaces/jboss-eap-quickstarts/buildconfigs/helloworld-pipeline/webhooks/kontinu8/github Problem is, GitHub isn't liking it. :) ? "We had a problem connecting to the server. The most common problem with these types of errors is a misconfigured SSL certificate. Try verifying your certificate chain." Short of installing a proper certificate on this server and giving it a domain, etc; any bright ideas to clear this up so we can get autobuild on push events? S, ALR -- Red Hat Developer Programs Architecture @ALRubinger _______________________________________________ kontinuity-dev-public mailing list kontinuity-dev-public at redhat.com https://www.redhat.com/mailman/listinfo/kontinuity-dev-public -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Sat Jun 25 05:36:33 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Sat, 25 Jun 2016 01:36:33 -0400 Subject: [kontinuity-dev-public] Webhooks to our remote OpenShift instance In-Reply-To: References: Message-ID: Nevermind, I got it. We can add an "insecure_ssl" field into the config when we create the webhook: https://developer.github.com/v3/repos/hooks/ So: https://github.com/redhat-kontinuity/catapult/issues/142 S, ALR On Sat, Jun 25, 2016 at 1:28 AM, Andrew Lee Rubinger wrote: > So we've equipped Catapult with the ability to add a webhook to the GitHub > repo so it kicks OpenShift when there's a push (yay!). > > It looks a little like: > > > https://147.75.202.53:8443/oapi/v1/namespaces/jboss-eap-quickstarts/buildconfigs/helloworld-pipeline/webhooks/kontinu8/github > > Problem is, GitHub isn't liking it. :) > > "We had a problem connecting to the server. The most common problem with > these types of errors is a misconfigured SSL certificate. Try verifying > your certificate chain." > > Short of installing a proper certificate on this server and giving it a > domain, etc; any bright ideas to clear this up so we can get autobuild on > push events? > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Sat Jun 25 05:38:41 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Sat, 25 Jun 2016 01:38:41 -0400 Subject: [kontinuity-dev-public] Webhooks to our remote OpenShift instance In-Reply-To: References: Message-ID: Ha, you beat me to the "send" button :) And of COURSE we should do it. This is what I call a problem for Future Us: https://github.com/redhat-kontinuity/catapult/issues/143 S, ALR On Sat, Jun 25, 2016 at 1:36 AM, Andrew Block wrote: > The GitHub API [1] supports setting an insecure flag to disable SSL > validation in web hooks. The java client does allow you to configure this > setting in the config field [2]. > > Whether we should allow that is another story, but it is an option. > > - Andy > > [1] - https://developer.github.com/v3/repos/hooks/ > [2] - > https://github.com/redhat-kontinuity/catapult/blob/master/services/github-service-impl/src/main/java/org/kontinuity/catapult/service/github/impl/kohsuke/KohsukeGitHubServiceImpl.java#L184-L186 > > > > Andrew Block > Red Hat Consulting > 101 N. Wacker, Suite 150 > Chicago, IL 60606 > andrew.block at redhat.com | m. (716) 870-2408 > > On June 25, 2016 at 12:28:54 AM, Andrew Lee Rubinger (alr at redhat.com) > wrote: > > So we've equipped Catapult with the ability to add a webhook to the GitHub > repo so it kicks OpenShift when there's a push (yay!). > > It looks a little like: > > > https://147.75.202.53:8443/oapi/v1/namespaces/jboss-eap-quickstarts/buildconfigs/helloworld-pipeline/webhooks/kontinu8/github > > Problem is, GitHub isn't liking it. :) > > "We had a problem connecting to the server. The most common problem with > these types of errors is a misconfigured SSL certificate. Try verifying > your certificate chain." > > Short of installing a proper certificate on this server and giving it a > domain, etc; any bright ideas to clear this up so we can get autobuild on > push events? > > S, > ALR > > -- > Red Hat Developer Programs Architecture > @ALRubinger > _______________________________________________ > kontinuity-dev-public mailing list > kontinuity-dev-public at redhat.com > https://www.redhat.com/mailman/listinfo/kontinuity-dev-public > > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From alr at redhat.com Sat Jun 25 08:18:38 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Sat, 25 Jun 2016 04:18:38 -0400 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely In-Reply-To: References: Message-ID: Played around a bunch and I think it's this: The build is triggered before Jenkins is up and running. Is there a more appropriate event we can use as a trigger like "Jenkins probe succeeds"? S, ALR On Fri, Jun 24, 2016 at 10:57 PM, Ben Parees wrote: > > On Jun 24, 2016 10:30 PM, "Andrew Lee Rubinger" wrote: > > > > Hi all: > > > > So with Ben's new changes in place (thanks!) the k8s namespace is set in > Jenkins kubernetes-plugin: > > > > https://ibin.co/2lpVENzMzOGm.png > > > > This gets us an OpenShift build triggered properly, but it waits > indefinitely without ever kicking off the build in Jenkins: > > > > https://ibin.co/2lpVJNi1p37O.png > > > > Manually triggering a second build in OpenShift calls into Jenkins > properly and completes as expected. > > > > So...race condition? Tries to kick off a build before namespace is set? > > No, the namespace is set in the jenkins config.xml before jenkins after > starts, so it's not that. (also it only affects slave launching, not > pipeline starting) But what it is i'm not sure. The first build is > created/started by you instantiating your template? I think it's a question > for jimmi as to why it doesn't see the new openshift build and start the > pipeline in response. > > > > > S, > > ALR > > > > -- > > Red Hat Developer Programs Architecture > > @ALRubinger > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From bparees at redhat.com Sat Jun 25 13:25:10 2016 From: bparees at redhat.com (Ben Parees) Date: Sat, 25 Jun 2016 09:25:10 -0400 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely In-Reply-To: References: Message-ID: On Jun 25, 2016 4:19 AM, "Andrew Lee Rubinger" wrote: > > Played around a bunch and I think it's this: The build is triggered before Jenkins is up and running. The sync plugin ought not care, when it comes up it should see the build in the new state and process it. Please open an issue against the sync plugin repo and tag Jimmi. (but we probably won't be rebuilding the image during summit week even if we get a fix) > > Is there a more appropriate event we can use as a trigger like "Jenkins probe succeeds"? So your template is, upon instantiation, kicking off your pipeline build? (presumably because you have a ConfigChangeTrigger on your pipeline buildconfig definition) You could remove that trigger and just manually start the build from openshift once you know jenkins is up, as a workaround. > > S, > ALR > > On Fri, Jun 24, 2016 at 10:57 PM, Ben Parees wrote: >> >> >> On Jun 24, 2016 10:30 PM, "Andrew Lee Rubinger" wrote: >> > >> > Hi all: >> > >> > So with Ben's new changes in place (thanks!) the k8s namespace is set in Jenkins kubernetes-plugin: >> > >> > https://ibin.co/2lpVENzMzOGm.png >> > >> > This gets us an OpenShift build triggered properly, but it waits indefinitely without ever kicking off the build in Jenkins: >> > >> > https://ibin.co/2lpVJNi1p37O.png >> > >> > Manually triggering a second build in OpenShift calls into Jenkins properly and completes as expected. >> > >> > So...race condition? Tries to kick off a build before namespace is set? >> >> No, the namespace is set in the jenkins config.xml before jenkins after starts, so it's not that. (also it only affects slave launching, not pipeline starting) But what it is i'm not sure. The first build is created/started by you instantiating your template? I think it's a question for jimmi as to why it doesn't see the new openshift build and start the pipeline in response. >> >> > >> > S, >> > ALR >> > >> > -- >> > Red Hat Developer Programs Architecture >> > @ALRubinger > > > > > -- > Red Hat Developer Programs Architecture > @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Sun Jun 26 18:32:32 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Sun, 26 Jun 2016 20:32:32 +0200 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely In-Reply-To: References: Message-ID: <1466965952.30610.7.camel@redhat.com> On So, 2016-06-25 at 09:25 -0400, Ben Parees wrote: > You could remove that trigger and just manually start the build from > > On Jun 25, 2016 4:19 AM, "Andrew Lee Rubinger" > wrote: > > > > Played around a bunch and I think it's this: The build is triggered > before Jenkins is up and running. > The sync plugin ought not care, when it comes up it should see the > build in the new state and process it. Please open an issue against > the sync plugin repo and tag Jimmi. (but we probably won't be > rebuilding the image during summit week even if we get a fix)? > > > > Is there a more appropriate event we can use as a trigger like > "Jenkins probe succeeds"? > So your template is, upon instantiation, kicking off your pipeline > build? (presumably because you have a ConfigChangeTrigger on your > pipeline buildconfig definition)openshift once you know jenkins is > up, as a workaround. Andy, I have disabled the automatic trigger for the first build as Ben suggested. Better to run it manually then having to cancel it and do the same. https://github.com/redhat-kontinuity/jboss-eap- quickstarts/commit/09eb555f515dc38844a012b18ade5efa319986d8 > > > > > S, > > ALR > > > > On Fri, Jun 24, 2016 at 10:57 PM, Ben Parees > wrote: > >> > >> > >> On Jun 24, 2016 10:30 PM, "Andrew Lee Rubinger" > wrote: > >> > > >> > Hi all: > >> > > >> > So with Ben's new changes in place (thanks!) the k8s namespace > is set in Jenkins kubernetes-plugin: > >> > > >> > ??https://ibin.co/2lpVENzMzOGm.png > >> > > >> > This gets us an OpenShift build triggered properly, but it waits > indefinitely without ever kicking off the build in Jenkins: > >> > > >> > ??https://ibin.co/2lpVJNi1p37O.png > >> > > >> > Manually triggering a second build in OpenShift calls into > Jenkins properly and completes as expected. > >> > > >> > So...race condition?? Tries to kick off a build before namespace > is set? > >> > >> No, the namespace is set in the jenkins config.xml before jenkins > after starts, so it's not that. (also it only affects slave > launching, not pipeline starting) But what it is i'm not sure.? The > first build is created/started by you instantiating your template? I > think it's a question for jimmi as to why it doesn't see the new > openshift build and start the pipeline in response. > >> > >> > > >> > S, > >> > ALR > >> > > >> > --? > >> > Red Hat Developer Programs Architecture > >> > @ALRubinger > > > > > > > > > > --? > > Red Hat Developer Programs Architecture > > @ALRubinger From alr at redhat.com Sun Jun 26 20:13:07 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Sun, 26 Jun 2016 13:13:07 -0700 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely In-Reply-To: <1466965952.30610.7.camel@redhat.com> References: <1466965952.30610.7.camel@redhat.com> Message-ID: On Sun, Jun 26, 2016 at 11:32 AM, Tomas Nozicka wrote: > On So, 2016-06-25 at 09:25 -0400, Ben Parees wrote: > > You could remove that trigger and just manually start the build from > > > > > On Jun 25, 2016 4:19 AM, "Andrew Lee Rubinger" > > wrote: > > > > > > Played around a bunch and I think it's this: The build is triggered > > before Jenkins is up and running. > > The sync plugin ought not care, when it comes up it should see the > > build in the new state and process it. Please open an issue against > > the sync plugin repo and tag Jimmi. (but we probably won't be > > rebuilding the image during summit week even if we get a fix) > > > > > > Is there a more appropriate event we can use as a trigger like > > "Jenkins probe succeeds"? > > So your template is, upon instantiation, kicking off your pipeline > > build? (presumably because you have a ConfigChangeTrigger on your > > pipeline buildconfig definition)openshift once you know jenkins is > > up, as a workaround. > Andy, I have disabled the automatic trigger for the first build as Ben > suggested. Better to run it manually then having to cancel it and do > the same. > https://github.com/redhat-kontinuity/jboss-eap- > quickstarts/commit/09eb555f515dc38844a012b18ade5efa319986d8 > OK, I agree; thank you! Mind doing the same for the kitchensink example (and the upcoming Node one from Rupali)? S, ALR > > > > > > > > > S, > > > ALR > > > > > > On Fri, Jun 24, 2016 at 10:57 PM, Ben Parees > > wrote: > > >> > > >> > > >> On Jun 24, 2016 10:30 PM, "Andrew Lee Rubinger" > > wrote: > > >> > > > >> > Hi all: > > >> > > > >> > So with Ben's new changes in place (thanks!) the k8s namespace > > is set in Jenkins kubernetes-plugin: > > >> > > > >> > https://ibin.co/2lpVENzMzOGm.png > > >> > > > >> > This gets us an OpenShift build triggered properly, but it waits > > indefinitely without ever kicking off the build in Jenkins: > > >> > > > >> > https://ibin.co/2lpVJNi1p37O.png > > >> > > > >> > Manually triggering a second build in OpenShift calls into > > Jenkins properly and completes as expected. > > >> > > > >> > So...race condition? Tries to kick off a build before namespace > > is set? > > >> > > >> No, the namespace is set in the jenkins config.xml before jenkins > > after starts, so it's not that. (also it only affects slave > > launching, not pipeline starting) But what it is i'm not sure. The > > first build is created/started by you instantiating your template? I > > think it's a question for jimmi as to why it doesn't see the new > > openshift build and start the pipeline in response. > > >> > > >> > > > >> > S, > > >> > ALR > > >> > > > >> > -- > > >> > Red Hat Developer Programs Architecture > > >> > @ALRubinger > > > > > > > > > > > > > > > -- > > > Red Hat Developer Programs Architecture > > > @ALRubinger > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Sun Jun 26 20:17:21 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Sun, 26 Jun 2016 22:17:21 +0200 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely In-Reply-To: References: <1466965952.30610.7.camel@redhat.com> Message-ID: <1466972241.30610.21.camel@redhat.com> On Ne, 2016-06-26 at 13:13 -0700, Andrew Lee Rubinger wrote: > > > On Sun, Jun 26, 2016 at 11:32 AM, Tomas Nozicka > wrote: > > On So, 2016-06-25 at 09:25 -0400, Ben Parees wrote: > > > You could remove that trigger and just manually start the build > > from > > > > > > > > On Jun 25, 2016 4:19 AM, "Andrew Lee Rubinger" > > > wrote: > > > > > > > > Played around a bunch and I think it's this: The build is > > triggered > > > before Jenkins is up and running. > > > The sync plugin ought not care, when it comes up it should see > > the > > > build in the new state and process it. Please open an issue > > against > > > the sync plugin repo and tag Jimmi. (but we probably won't be > > > rebuilding the image during summit week even if we get a fix)? > > > > > > > > Is there a more appropriate event we can use as a trigger like > > > "Jenkins probe succeeds"? > > > So your template is, upon instantiation, kicking off your > > pipeline > > > build? (presumably because you have a ConfigChangeTrigger on your > > > pipeline buildconfig definition)openshift once you know jenkins > > is > > > up, as a workaround. > > Andy, I have disabled the automatic trigger for the first build as > > Ben > > suggested. Better to run it manually then having to cancel it and > > do > > the same. > > https://github.com/redhat-kontinuity/jboss-eap- > > quickstarts/commit/09eb555f515dc38844a012b18ade5efa319986d8 > OK, I agree; thank you!? Mind doing the same for the kitchensink > example (and the upcoming Node one from Rupali)? Sure, I will update those tomorrow with the fix for delayed deployment. > > S, > ALR > ? > > > > > > > > > > > > > S, > > > > ALR > > > > > > > > On Fri, Jun 24, 2016 at 10:57 PM, Ben Parees > m> > > > wrote: > > > >> > > > >> > > > >> On Jun 24, 2016 10:30 PM, "Andrew Lee Rubinger" > m> > > > wrote: > > > >> > > > > >> > Hi all: > > > >> > > > > >> > So with Ben's new changes in place (thanks!) the k8s > > namespace > > > is set in Jenkins kubernetes-plugin: > > > >> > > > > >> > ??https://ibin.co/2lpVENzMzOGm.png > > > >> > > > > >> > This gets us an OpenShift build triggered properly, but it > > waits > > > indefinitely without ever kicking off the build in Jenkins: > > > >> > > > > >> > ??https://ibin.co/2lpVJNi1p37O.png > > > >> > > > > >> > Manually triggering a second build in OpenShift calls into > > > Jenkins properly and completes as expected. > > > >> > > > > >> > So...race condition?? Tries to kick off a build before > > namespace > > > is set? > > > >> > > > >> No, the namespace is set in the jenkins config.xml before > > jenkins > > > after starts, so it's not that. (also it only affects slave > > > launching, not pipeline starting) But what it is i'm not sure.? > > The > > > first build is created/started by you instantiating your > > template? I > > > think it's a question for jimmi as to why it doesn't see the new > > > openshift build and start the pipeline in response. > > > >> > > > >> > > > > >> > S, > > > >> > ALR > > > >> > > > > >> > --? > > > >> > Red Hat Developer Programs Architecture > > > >> > @ALRubinger > > > > > > > > > > > > > > > > > > > > --? > > > > Red Hat Developer Programs Architecture > > > > @ALRubinger > > > > > --? > Red Hat Developer Programs Architecture > @ALRubinger From alr at redhat.com Sun Jun 26 20:20:29 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Sun, 26 Jun 2016 13:20:29 -0700 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely In-Reply-To: <1466972241.30610.21.camel@redhat.com> References: <1466965952.30610.7.camel@redhat.com> <1466972241.30610.21.camel@redhat.com> Message-ID: On Sun, Jun 26, 2016 at 1:17 PM, Tomas Nozicka wrote: > On Ne, 2016-06-26 at 13:13 -0700, Andrew Lee Rubinger wrote: > > > > > > On Sun, Jun 26, 2016 at 11:32 AM, Tomas Nozicka > > wrote: > > > On So, 2016-06-25 at 09:25 -0400, Ben Parees wrote: > > > > You could remove that trigger and just manually start the build > > > from > > > > > > > > > > > On Jun 25, 2016 4:19 AM, "Andrew Lee Rubinger" > > > > wrote: > > > > > > > > > > Played around a bunch and I think it's this: The build is > > > triggered > > > > before Jenkins is up and running. > > > > The sync plugin ought not care, when it comes up it should see > > > the > > > > build in the new state and process it. Please open an issue > > > against > > > > the sync plugin repo and tag Jimmi. (but we probably won't be > > > > rebuilding the image during summit week even if we get a fix) > > > > > > > > > > Is there a more appropriate event we can use as a trigger like > > > > "Jenkins probe succeeds"? > > > > So your template is, upon instantiation, kicking off your > > > pipeline > > > > build? (presumably because you have a ConfigChangeTrigger on your > > > > pipeline buildconfig definition)openshift once you know jenkins > > > is > > > > up, as a workaround. > > > Andy, I have disabled the automatic trigger for the first build as > > > Ben > > > suggested. Better to run it manually then having to cancel it and > > > do > > > the same. > > > https://github.com/redhat-kontinuity/jboss-eap- > > > quickstarts/commit/09eb555f515dc38844a012b18ade5efa319986d8 > > OK, I agree; thank you! Mind doing the same for the kitchensink > > example (and the upcoming Node one from Rupali)? > Sure, I will update those tomorrow with the fix for delayed deployment. > Thank you! BTW just ran through this demo with MartijnV (leader of the London JUG; CEO of jClarity) to really positive reviews, which I'm hoping is reflective of the type of response we can see to the features we've been building. :) Applying the two fixes you mention will basically eliminate the user experience hiccups from the demo and really keep the message on-point. S, ALR > > > > > S, > > ALR > > > > > > > > > > > > > > > > > > > S, > > > > > ALR > > > > > > > > > > On Fri, Jun 24, 2016 at 10:57 PM, Ben Parees > > m> > > > > wrote: > > > > >> > > > > >> > > > > >> On Jun 24, 2016 10:30 PM, "Andrew Lee Rubinger" > > m> > > > > wrote: > > > > >> > > > > > >> > Hi all: > > > > >> > > > > > >> > So with Ben's new changes in place (thanks!) the k8s > > > namespace > > > > is set in Jenkins kubernetes-plugin: > > > > >> > > > > > >> > https://ibin.co/2lpVENzMzOGm.png > > > > >> > > > > > >> > This gets us an OpenShift build triggered properly, but it > > > waits > > > > indefinitely without ever kicking off the build in Jenkins: > > > > >> > > > > > >> > https://ibin.co/2lpVJNi1p37O.png > > > > >> > > > > > >> > Manually triggering a second build in OpenShift calls into > > > > Jenkins properly and completes as expected. > > > > >> > > > > > >> > So...race condition? Tries to kick off a build before > > > namespace > > > > is set? > > > > >> > > > > >> No, the namespace is set in the jenkins config.xml before > > > jenkins > > > > after starts, so it's not that. (also it only affects slave > > > > launching, not pipeline starting) But what it is i'm not sure. > > > The > > > > first build is created/started by you instantiating your > > > template? I > > > > think it's a question for jimmi as to why it doesn't see the new > > > > openshift build and start the pipeline in response. > > > > >> > > > > >> > > > > > >> > S, > > > > >> > ALR > > > > >> > > > > > >> > -- > > > > >> > Red Hat Developer Programs Architecture > > > > >> > @ALRubinger > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > Red Hat Developer Programs Architecture > > > > > @ALRubinger > > > > > > > > > -- > > Red Hat Developer Programs Architecture > > @ALRubinger > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: From tnozicka at redhat.com Mon Jun 27 14:50:01 2016 From: tnozicka at redhat.com (Tomas Nozicka) Date: Mon, 27 Jun 2016 16:50:01 +0200 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely In-Reply-To: References: <1466965952.30610.7.camel@redhat.com> <1466972241.30610.21.camel@redhat.com> Message-ID: <1467039001.23428.2.camel@redhat.com> On Ne, 2016-06-26 at 13:20 -0700, Andrew Lee Rubinger wrote: > > > On Sun, Jun 26, 2016 at 1:17 PM, Tomas Nozicka > wrote: > > On Ne, 2016-06-26 at 13:13 -0700, Andrew Lee Rubinger wrote: > > > > > > > > > On Sun, Jun 26, 2016 at 11:32 AM, Tomas Nozicka > com> > > > wrote: > > > > On So, 2016-06-25 at 09:25 -0400, Ben Parees wrote: > > > > > You could remove that trigger and just manually start the > > build > > > > from > > > > > > > > > > > > > > On Jun 25, 2016 4:19 AM, "Andrew Lee Rubinger" > m> > > > > > wrote: > > > > > > > > > > > > Played around a bunch and I think it's this: The build is > > > > triggered > > > > > before Jenkins is up and running. > > > > > The sync plugin ought not care, when it comes up it should > > see > > > > the > > > > > build in the new state and process it. Please open an issue > > > > against > > > > > the sync plugin repo and tag Jimmi. (but we probably won't be > > > > > rebuilding the image during summit week even if we get a > > fix)? > > > > > > > > > > > > Is there a more appropriate event we can use as a trigger > > like > > > > > "Jenkins probe succeeds"? > > > > > So your template is, upon instantiation, kicking off your > > > > pipeline > > > > > build? (presumably because you have a ConfigChangeTrigger on > > your > > > > > pipeline buildconfig definition)openshift once you know > > jenkins > > > > is > > > > > up, as a workaround. > > > > Andy, I have disabled the automatic trigger for the first build > > as > > > > Ben > > > > suggested. Better to run it manually then having to cancel it > > and > > > > do > > > > the same. > > > > https://github.com/redhat-kontinuity/jboss-eap- > > > > quickstarts/commit/09eb555f515dc38844a012b18ade5efa319986d8 > > > OK, I agree; thank you!? Mind doing the same for the kitchensink > > > example (and the upcoming Node one from Rupali)? > > Sure, I will update those tomorrow with the fix for delayed > > deployment. > Thank you! > > BTW just ran through this demo with MartijnV (leader of the London > JUG; CEO of jClarity) to really positive reviews, which I'm hoping is > reflective of the type of response we can see to the features we've > been building. :) That's awesome! > Applying the two fixes you mention will basically eliminate the user > experience hiccups from the demo and really keep the message on- > point. It is all fixed now. + I put PR for Node.js app in place. > > S, > ALR? > > > > > > > > S, > > > ALR > > > ? > > > > > > > > > > > > > > > > > > > > > S, > > > > > > ALR > > > > > > > > > > > > On Fri, Jun 24, 2016 at 10:57 PM, Ben Parees > t.co > > > > m> > > > > > wrote: > > > > > >> > > > > > >> > > > > > >> On Jun 24, 2016 10:30 PM, "Andrew Lee Rubinger" > t.co > > > > m> > > > > > wrote: > > > > > >> > > > > > > >> > Hi all: > > > > > >> > > > > > > >> > So with Ben's new changes in place (thanks!) the k8s > > > > namespace > > > > > is set in Jenkins kubernetes-plugin: > > > > > >> > > > > > > >> > ??https://ibin.co/2lpVENzMzOGm.png > > > > > >> > > > > > > >> > This gets us an OpenShift build triggered properly, but > > it > > > > waits > > > > > indefinitely without ever kicking off the build in Jenkins: > > > > > >> > > > > > > >> > ??https://ibin.co/2lpVJNi1p37O.png > > > > > >> > > > > > > >> > Manually triggering a second build in OpenShift calls > > into > > > > > Jenkins properly and completes as expected. > > > > > >> > > > > > > >> > So...race condition?? Tries to kick off a build before > > > > namespace > > > > > is set? > > > > > >> > > > > > >> No, the namespace is set in the jenkins config.xml before > > > > jenkins > > > > > after starts, so it's not that. (also it only affects slave > > > > > launching, not pipeline starting) But what it is i'm not > > sure.? > > > > The > > > > > first build is created/started by you instantiating your > > > > template? I > > > > > think it's a question for jimmi as to why it doesn't see the > > new > > > > > openshift build and start the pipeline in response. > > > > > >> > > > > > >> > > > > > > >> > S, > > > > > >> > ALR > > > > > >> > > > > > > >> > --? > > > > > >> > Red Hat Developer Programs Architecture > > > > > >> > @ALRubinger > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > --? > > > > > > Red Hat Developer Programs Architecture > > > > > > @ALRubinger > > > > > > > > > > > > > --? > > > Red Hat Developer Programs Architecture > > > @ALRubinger > > > > > --? > Red Hat Developer Programs Architecture > @ALRubinger From alr at redhat.com Tue Jun 28 21:20:50 2016 From: alr at redhat.com (Andrew Lee Rubinger) Date: Tue, 28 Jun 2016 17:20:50 -0400 Subject: [kontinuity-dev-public] k8s Namespace autoset, but build stalls indefinitely In-Reply-To: <1467039001.23428.2.camel@redhat.com> References: <1466965952.30610.7.camel@redhat.com> <1466972241.30610.21.camel@redhat.com> <1467039001.23428.2.camel@redhat.com> Message-ID: On Mon, Jun 27, 2016 at 10:50 AM, Tomas Nozicka wrote: > On Ne, 2016-06-26 at 13:20 -0700, Andrew Lee Rubinger wrote: > > > > > > On Sun, Jun 26, 2016 at 1:17 PM, Tomas Nozicka > > wrote: > > > On Ne, 2016-06-26 at 13:13 -0700, Andrew Lee Rubinger wrote: > > > > > > > > > > > > On Sun, Jun 26, 2016 at 11:32 AM, Tomas Nozicka > > com> > > > > wrote: > > > > > On So, 2016-06-25 at 09:25 -0400, Ben Parees wrote: > > > > > > You could remove that trigger and just manually start the > > > build > > > > > from > > > > > > > > > > > > > > > > > On Jun 25, 2016 4:19 AM, "Andrew Lee Rubinger" > > m> > > > > > > wrote: > > > > > > > > > > > > > > Played around a bunch and I think it's this: The build is > > > > > triggered > > > > > > before Jenkins is up and running. > > > > > > The sync plugin ought not care, when it comes up it should > > > see > > > > > the > > > > > > build in the new state and process it. Please open an issue > > > > > against > > > > > > the sync plugin repo and tag Jimmi. (but we probably won't be > > > > > > rebuilding the image during summit week even if we get a > > > fix) > > > > > > > > > > > > > > Is there a more appropriate event we can use as a trigger > > > like > > > > > > "Jenkins probe succeeds"? > > > > > > So your template is, upon instantiation, kicking off your > > > > > pipeline > > > > > > build? (presumably because you have a ConfigChangeTrigger on > > > your > > > > > > pipeline buildconfig definition)openshift once you know > > > jenkins > > > > > is > > > > > > up, as a workaround. > > > > > Andy, I have disabled the automatic trigger for the first build > > > as > > > > > Ben > > > > > suggested. Better to run it manually then having to cancel it > > > and > > > > > do > > > > > the same. > > > > > https://github.com/redhat-kontinuity/jboss-eap- > > > > > quickstarts/commit/09eb555f515dc38844a012b18ade5efa319986d8 > > > > OK, I agree; thank you! Mind doing the same for the kitchensink > > > > example (and the upcoming Node one from Rupali)? > > > Sure, I will update those tomorrow with the fix for delayed > > > deployment. > > Thank you! > > > > BTW just ran through this demo with MartijnV (leader of the London > > JUG; CEO of jClarity) to really positive reviews, which I'm hoping is > > reflective of the type of response we can see to the features we've > > been building. :) > That's awesome! > > > Applying the two fixes you mention will basically eliminate the user > > experience hiccups from the demo and really keep the message on- > > point. > It is all fixed now. + I put PR for Node.js app in place. > Tested and working and fantastic. :) S, ALR > > > > > S, > > ALR > > > > > > > > > > > S, > > > > ALR > > > > > > > > > > > > > > > > > > > > > > > > > > > > > S, > > > > > > > ALR > > > > > > > > > > > > > > On Fri, Jun 24, 2016 at 10:57 PM, Ben Parees > > t.co > > > > > m> > > > > > > wrote: > > > > > > >> > > > > > > >> > > > > > > >> On Jun 24, 2016 10:30 PM, "Andrew Lee Rubinger" > > t.co > > > > > m> > > > > > > wrote: > > > > > > >> > > > > > > > >> > Hi all: > > > > > > >> > > > > > > > >> > So with Ben's new changes in place (thanks!) the k8s > > > > > namespace > > > > > > is set in Jenkins kubernetes-plugin: > > > > > > >> > > > > > > > >> > https://ibin.co/2lpVENzMzOGm.png > > > > > > >> > > > > > > > >> > This gets us an OpenShift build triggered properly, but > > > it > > > > > waits > > > > > > indefinitely without ever kicking off the build in Jenkins: > > > > > > >> > > > > > > > >> > https://ibin.co/2lpVJNi1p37O.png > > > > > > >> > > > > > > > >> > Manually triggering a second build in OpenShift calls > > > into > > > > > > Jenkins properly and completes as expected. > > > > > > >> > > > > > > > >> > So...race condition? Tries to kick off a build before > > > > > namespace > > > > > > is set? > > > > > > >> > > > > > > >> No, the namespace is set in the jenkins config.xml before > > > > > jenkins > > > > > > after starts, so it's not that. (also it only affects slave > > > > > > launching, not pipeline starting) But what it is i'm not > > > sure. > > > > > The > > > > > > first build is created/started by you instantiating your > > > > > template? I > > > > > > think it's a question for jimmi as to why it doesn't see the > > > new > > > > > > openshift build and start the pipeline in response. > > > > > > >> > > > > > > >> > > > > > > > >> > S, > > > > > > >> > ALR > > > > > > >> > > > > > > > >> > -- > > > > > > >> > Red Hat Developer Programs Architecture > > > > > > >> > @ALRubinger > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > Red Hat Developer Programs Architecture > > > > > > > @ALRubinger > > > > > > > > > > > > > > > > > -- > > > > Red Hat Developer Programs Architecture > > > > @ALRubinger > > > > > > > > > -- > > Red Hat Developer Programs Architecture > > @ALRubinger > -- Red Hat Developer Programs Architecture @ALRubinger -------------- next part -------------- An HTML attachment was scrubbed... URL: