From lzap at redhat.com Thu Dec 1 12:05:19 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 1 Dec 2011 13:05:19 +0100 Subject: [katello-devel] Passing separate arguments for a single shadow to text-shadow is deprecated Message-ID: <20111201120519.GC4696@lzapx.brq.redhat.com> Hello, I have the following warnings while compiling CSS. I am really not sure what they mean and if it is dangerous, FYI. http://pastebin.com/FVRmV7Dg -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Thu Dec 1 13:05:05 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 1 Dec 2011 14:05:05 +0100 Subject: [katello-devel] QQ on Ueber Cert stuff In-Reply-To: <55ef597d-2720-43dd-91a5-e9ebe90ebb2b@zmail14.collab.prod.int.phx2.redhat.com> References: <3d009a54-bdff-4629-8109-106054a463ef@zmail14.collab.prod.int.phx2.redhat.com> <55ef597d-2720-43dd-91a5-e9ebe90ebb2b@zmail14.collab.prod.int.phx2.redhat.com> Message-ID: <20111201130505.GE4696@lzapx.brq.redhat.com> I don't think Ueber certs need to be downloaded to a client. They stay on the Katello server. LZ On Wed, Nov 30, 2011 at 06:46:49PM -0500, Partha Aji wrote: > > So I was trying to implement the Ueber cert functionality in the UI and was wondering the preferred way of handling this. For those who don't know what Ueber Certs are go to https://fedorahosted.org/katello/wiki/GuideDebugCertificates > > I was thinking along the lines of adding Generate Cert and Download Cert buttons on Organization Details and having it do what the API does at present. > > My question is should we just make the user download the pkcs12 file instead of separately downloading the pem files and then having to the run "openssl pkcs12 -keypbe ...". If yes, then my next question is what do I do about the password that is required to generate the pkcs12 ? (I guess I can ask the user for a password when they ask me to generate the certificate) > > The alternative is ofcourse offer 2 download links , one to download the key pem file and the other to download the cert pem. > > > Partha > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From bkearney at redhat.com Thu Dec 1 13:16:05 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Thu, 01 Dec 2011 08:16:05 -0500 Subject: [katello-devel] QQ on Ueber Cert stuff In-Reply-To: <20111201130505.GE4696@lzapx.brq.redhat.com> References: <3d009a54-bdff-4629-8109-106054a463ef@zmail14.collab.prod.int.phx2.redhat.com> <55ef597d-2720-43dd-91a5-e9ebe90ebb2b@zmail14.collab.prod.int.phx2.redhat.com> <20111201130505.GE4696@lzapx.brq.redhat.com> Message-ID: <4ED77E15.2030002@redhat.com> On 12/01/2011 08:05 AM, Lukas Zapletal wrote: > I don't think Ueber certs need to be downloaded to a client. They stay > on the Katello server. > I tend to agree this is not a common case. I would leave them as seperate PEM files.. that way i can have different passwords on my pkcs12. KISS for now. -- bk From ehelms at redhat.com Thu Dec 1 13:22:34 2011 From: ehelms at redhat.com (Eric Helms) Date: Thu, 01 Dec 2011 08:22:34 -0500 (EST) Subject: [katello-devel] Passing separate arguments for a single shadow to text-shadow is deprecated In-Reply-To: <20111201120519.GC4696@lzapx.brq.redhat.com> Message-ID: Compass, which we use to help manage and write our CSS, has undergone small API changes as we have updated that gem. Don't think it is dangerous, but since we have a story this sprint related to making the UI more consistent I will look into fixing these deprecated methods. Thanks, Eric ----- Original Message ----- From: "Lukas Zapletal" To: katello-devel at redhat.com Sent: Thursday, December 1, 2011 7:05:19 AM Subject: [katello-devel] Passing separate arguments for a single shadow to text-shadow is deprecated Hello, I have the following warnings while compiling CSS. I am really not sure what they mean and if it is dangerous, FYI. http://pastebin.com/FVRmV7Dg -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno _______________________________________________ katello-devel mailing list katello-devel at redhat.com https://www.redhat.com/mailman/listinfo/katello-devel From lzap at redhat.com Thu Dec 1 13:43:33 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 1 Dec 2011 14:43:33 +0100 Subject: [katello-devel] QQ on Ueber Cert stuff In-Reply-To: <4ED77E15.2030002@redhat.com> References: <3d009a54-bdff-4629-8109-106054a463ef@zmail14.collab.prod.int.phx2.redhat.com> <55ef597d-2720-43dd-91a5-e9ebe90ebb2b@zmail14.collab.prod.int.phx2.redhat.com> <20111201130505.GE4696@lzapx.brq.redhat.com> <4ED77E15.2030002@redhat.com> Message-ID: <20111201134333.GG4696@lzapx.brq.redhat.com> On Thu, Dec 01, 2011 at 08:16:05AM -0500, Bryan Kearney wrote: > On 12/01/2011 08:05 AM, Lukas Zapletal wrote: > >I don't think Ueber certs need to be downloaded to a client. They stay > >on the Katello server. > > > > I tend to agree this is not a common case. I would leave them as > seperate PEM files.. that way i can have different passwords on my > pkcs12. KISS for now. > My understanding is we generate them and handing over to Image Factory. They can also be used to browse content, but for this we could provide a cli command. Do we need it in the UI? -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Thu Dec 1 17:00:19 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 1 Dec 2011 18:00:19 +0100 Subject: [katello-devel] Manifest cat utility Message-ID: <20111201170019.GK4696@lzapx.brq.redhat.com> Guys, I wrote a simple script that prints useful information from manifest ZIP file. Its here: git:scripts/catmf Be sure you symlink it to your $HOME/bin dir, very useful. Example session (fake manifest): $ catmf fake-manifest.zip Product: rand-19248 (rand-99053) Product: rand-64228 (rand-66393) Entitlement (ff808081336da11101336da917760007): Product: rand-64228 (rand-66393) Pool Id: ff808081336da11101336da917760007 Quantity: 30 Contract: Account: 76534 Ends: 2025-05-29T00:00:00.000+0000 Entitlement (ff808081336da11101336da917fe0009): Product: rand-19248 (rand-99053) Pool Id: ff808081336da11101336da917fe0009 Quantity: 20 Contract: Account: 12345 Ends: 2025-05-29T00:00:00.000+0000 Real one: http://pastebin.test.redhat.com/69673 -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From mmccune at redhat.com Thu Dec 1 17:04:32 2011 From: mmccune at redhat.com (Mike McCune) Date: Thu, 01 Dec 2011 09:04:32 -0800 Subject: [katello-devel] Manifest cat utility In-Reply-To: <20111201170019.GK4696@lzapx.brq.redhat.com> References: <20111201170019.GK4696@lzapx.brq.redhat.com> Message-ID: <4ED7B3A0.5020907@redhat.com> On 12/01/2011 09:00 AM, Lukas Zapletal wrote: > Guys, > > I wrote a simple script that prints useful information from manifest ZIP > file. Its here: > > git:scripts/catmf > nice! super useful From jrist at redhat.com Thu Dec 1 17:49:42 2011 From: jrist at redhat.com (Jason Rist) Date: Thu, 01 Dec 2011 10:49:42 -0700 Subject: [katello-devel] Passing separate arguments for a single shadow to text-shadow is deprecated In-Reply-To: References: <20111201120519.GC4696@lzapx.brq.redhat.com> Message-ID: <4ED7BE36.1000301@redhat.com> On Thu 01 Dec 2011 06:22:34 AM MST, Eric Helms wrote: > Compass, which we use to help manage and write our CSS, has undergone small API changes as we have updated that gem. Don't think it is dangerous, but since we have a story this sprint related to making the UI more consistent I will look into fixing these deprecated methods. > > Thanks, > Eric > > ----- Original Message ----- > From: "Lukas Zapletal" > To: katello-devel at redhat.com > Sent: Thursday, December 1, 2011 7:05:19 AM > Subject: [katello-devel] Passing separate arguments for a single shadow to text-shadow is deprecated > > Hello, > > I have the following warnings while compiling CSS. I am really not sure > what they mean and if it is dangerous, FYI. > > http://pastebin.com/FVRmV7Dg > I could have sworn I fixed those in master as part of the branding work. We will definitely fix those as we see them. -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From mmccune at redhat.com Thu Dec 1 21:19:41 2011 From: mmccune at redhat.com (Mike McCune) Date: Thu, 01 Dec 2011 13:19:41 -0800 Subject: [katello-devel] more console.log in our JavaScript Message-ID: <4ED7EF6D.3090106@redhat.com> these kill Firefox 3.6 dead: $ git grep console |grep -v jquery javascripts/gpg_key.js: console.log(data); javascripts/gpg_key.js: console.log('error'); javascripts/product_edit.js: console.log(result); javascripts/system_template.js: console.log(dist.id + "," + current.distribution) I looked to see if JSL could catch these and break the build but I didn't see a simple solution. Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From jrist at redhat.com Fri Dec 2 00:02:23 2011 From: jrist at redhat.com (Jason Rist) Date: Thu, 01 Dec 2011 17:02:23 -0700 Subject: [katello-devel] more console.log in our JavaScript In-Reply-To: <4ED7EF6D.3090106@redhat.com> References: <4ED7EF6D.3090106@redhat.com> Message-ID: <4ED8158F.9030006@redhat.com> On Thu 01 Dec 2011 02:19:41 PM MST, Mike McCune wrote: > these kill Firefox 3.6 dead: > > $ git grep console |grep -v jquery > javascripts/gpg_key.js: console.log(data); > javascripts/gpg_key.js: console.log('error'); > javascripts/product_edit.js: console.log(result); > javascripts/system_template.js: console.log(dist.id + "," + > current.distribution) > > I looked to see if JSL could catch these and break the build but I > didn't see a simple solution. > > Mike There is a slightly better way around this, by checking for the console: if ( typeof console !== "undefined" && console.error && console.warn ) { console.error("Ugh, this is an error."); console.log("Just loggin' some stuff."); } If you're going to do any logging to the console, please do the above. If you're using it for debugging, you're doing something wrong :) -J -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From lzap at redhat.com Fri Dec 2 13:49:27 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 2 Dec 2011 14:49:27 +0100 Subject: [katello-devel] Adding new list_permissions rake task Message-ID: <20111202134927.GF3395@lzapx.brq.redhat.com> Guys, I wrote a simple, yet powerful rake task: # rake list_permissions It generates a Trac-wiki compatible table of all our controllers and actions (defined in the current routes.rb) and for each action it finds corresponding permission from our rules hash. As I am not able to show source of our permissions lambdas in Ruby 1.8, I put a link into our gitweb for those who want to check particular rule out. Simply put, it shows us all actions that are unprotected. And for both UI and API. As I will work on API permissions this sprint I will be updating this page with the latest and greatest list_permissions output: https://fedorahosted.org/katello/wiki/PermissionRoutesCheck Feel free to update the page every time you work on permissions. Those with blank third column are unprotected right now. We should add this task to our list of pre-release bullets that should be always checked to prevent security flaws. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Fri Dec 2 15:27:32 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 2 Dec 2011 16:27:32 +0100 Subject: [katello-devel] Running katello-configure multiple times Message-ID: <20111202152732.GH3395@lzapx.brq.redhat.com> Hello, it seems few folks are trying to run katello-configure multiple times. I believe Rails seed script (or our "rake setup" task) do not support this, I would recommend to prevent this with a simple lock file. For me it shows: rake aborted! Validation failed: Name must be unique within an organization scope if I run it 2nd time or more. OR We could make our rake setup task idempotent. If I can remember correctly, we had this already. Opinions? -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From morazi at redhat.com Fri Dec 2 15:48:07 2011 From: morazi at redhat.com (Mike Orazi) Date: Fri, 02 Dec 2011 10:48:07 -0500 Subject: [katello-devel] Running katello-configure multiple times In-Reply-To: <20111202152732.GH3395@lzapx.brq.redhat.com> References: <20111202152732.GH3395@lzapx.brq.redhat.com> Message-ID: <1322840887.2402.12.camel@morazi-lt.usersys.redhat.com> On Fri, 2011-12-02 at 16:27 +0100, Lukas Zapletal wrote: > Hello, > > it seems few folks are trying to run katello-configure multiple times. > > I believe Rails seed script (or our "rake setup" task) do not support > this, I would recommend to prevent this with a simple lock file. > > For me it shows: > > rake aborted! > Validation failed: Name must be unique within an organization scope > > if I run it 2nd time or more. > > OR > > We could make our rake setup task idempotent. If I can remember > correctly, we had this already. > > Opinions? > iirc aeolus-configure currently uses a lock file -- it was just way easier at the time. If you end up making the rake task idempotent I'd love to hear about it, and would probably try to find a way to keep things consistent between the projects. Thanks, Mike From jortel at redhat.com Fri Dec 2 16:38:45 2011 From: jortel at redhat.com (Jeff Ortel) Date: Fri, 02 Dec 2011 10:38:45 -0600 Subject: [katello-devel] Pulp API mods needed for system event history In-Reply-To: <744af441-4693-4a93-89eb-222668bca9a5@zmail14.collab.prod.int.phx2.redhat.com> References: <744af441-4693-4a93-89eb-222668bca9a5@zmail14.collab.prod.int.phx2.redhat.com> Message-ID: <4ED8FF15.1020703@redhat.com> I assume "statuses" in (1-4) mean Task.state as defined within the Pulp tasking subsystem and exposed through our REST API. Is this correct? On 11/30/2011 03:09 PM, Partha Aji wrote: > > Brad and myself are working on a story to enable to one to see events that transpired on a System. Examples being things like Package Installs/Update/Remove, package profile request, auto heal updates etc. Before we model things in Katello I was wondering if the following stories were doable in pulp (and possibly in Candlepin). If the abilities listed below are already there in pulp or candlepin then please let me know. > > Things that we'd ideally need to be in pulp and probably in candlepin also > > 1) As a user I would like to get a list of statuses/tasks associated to a consumer. Pulp can provide this but currently does not model the association between consumers and tasks as a 1st class association. It can be derived from task arguments but since the consumer ID is potentially not positionally the same (depending on what the task is doing) this /may/ require modification to our tasking subsystem. Pulp's tasking subsystem does provide a good query mechanism that we currently don't expose through the REST API. If katello could query task status based on what the task was doing (I know, kind of wanky), supporting this query could be as simple as exposing through the REST API. > 2) As a user I would like to get a list of statuses based on a list of status id's If "status ids" means task IDs, this one is easy but would need some REST API enhancements to support passing multiple task IDs. > 3) As a user I would like to be able to get filtered list of tasks based on task states (like pending/completed/failed/cancelled/timed_out) Again, exposing more of our tasking (subsystem) query mechanism through the REST API, we could easily support this. > 4) As a user I would like the returned status data also send me back the parameters I passed in when I created the task. For example if I asked pulp to install a list of packages, I would like the status information for that task give me the name of the packages I asked it to install. > Currently the task (callable) arguments contains the parameters katello passed and the Task.result contains the list of (exact nevra) packages actually installed. However, for things like errata there is a disconnect between the API request (including parameters passed) and the task you get back. For example, what actually happens in pulp when you request errata install: 1) API call with errata IDs 2) Pulp server figures out which packages the errata references. 3) Creates a task to install/update those packages on the consumer. So, the task is actually a package install task so it's parameters will *not* contain the errata IDs but rather package names. The basic issue here is that we expose the raw (internal) representation of tasks through our REST API. We're aware of this and I've discussed with jconnor. We have some candidate solutions for this but requires more design. > > We'd have to tailor the Katello SystemTask status model differently based on what tasks get completed in the above list. Although pulp currently supports tasking and task querying and could support these things it may still make sense for katello to enhance it's modeling to support (1-4) for the following reasons: 1) Katello already stores pulp task IDs and associates to katello (operations, user requests,..) on consumers. 2) Task in pulp are to some degree an implementation detail in pulp and don't necessarily align with tasks (operations, user requests,..) in katello. They may not represent the same thing in both systems. 3) Same as #2 with regard to Katello and Candlepin. > > > Partha From lzap at redhat.com Fri Dec 2 17:08:30 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 2 Dec 2011 18:08:30 +0100 Subject: [katello-devel] Fully qualified domain name Message-ID: <20111202170830.GI3395@lzapx.brq.redhat.com> Hello, it seems that the seed error we see very often in Puppet occurs, when domainname is not set properly. Before running katello-configure please check the following things: 1) You have FQDN set $ hostname myhostname.mydomain 2) You have both in the /etc/hosts or your DNS resolves everything properly: $ ping -c1 localhost $ ping -c1 myhostname.mydomain I will add a check to the katello-configure script that will throw an error. More info at: https://bugzilla.redhat.com/show_bug.cgi?id=758712 Thank to gszasz and inecas for helping with tracking this ugly one. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Fri Dec 2 17:10:06 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 2 Dec 2011 18:10:06 +0100 Subject: [katello-devel] Running katello-configure multiple times In-Reply-To: <1322840887.2402.12.camel@morazi-lt.usersys.redhat.com> References: <20111202152732.GH3395@lzapx.brq.redhat.com> <1322840887.2402.12.camel@morazi-lt.usersys.redhat.com> Message-ID: <20111202171006.GJ3395@lzapx.brq.redhat.com> > iirc aeolus-configure currently uses a lock file -- it was just way > easier at the time. > I will add that as well on Monday. > If you end up making the rake task idempotent I'd love to hear about it, > and would probably try to find a way to keep things consistent between > the projects. I think its impossible to do that. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Mon Dec 5 14:40:30 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Mon, 5 Dec 2011 15:40:30 +0100 Subject: [katello-devel] Small repo with distribution Message-ID: <20111205144030.GF26396@lzapx.brq.redhat.com> Hey, it annoys me for a long time, so let me ask: Does anybody have a tip on a repo with a distribution that is small? Unfortunately Pulp does not provide to create a distribution. The only way to get it is to sync something. I am using Fedora for my testing. I did modify Tomas' ZOO repo and added some package groups, categories and files (empty iso file). But I was not able to find a simple way how to create a small repo with a distribution. http://lzap.fedorapeople.org/fakerepos/zoo2/ I'd be happy for any hints. I would love to include also distributions in our cli "smoke" tests, but it is not possible to sync Fedora each run. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From shughes at redhat.com Mon Dec 5 14:51:42 2011 From: shughes at redhat.com (Shannon Hughes) Date: Mon, 05 Dec 2011 09:51:42 -0500 Subject: [katello-devel] Small repo with distribution In-Reply-To: <20111205144030.GF26396@lzapx.brq.redhat.com> References: <20111205144030.GF26396@lzapx.brq.redhat.com> Message-ID: <4EDCDA7E.2050209@redhat.com> i used some of John Matthews small repos when testing sync error conditions. He might still have them up or can assist. On 12/05/2011 09:40 AM, Lukas Zapletal wrote: > Hey, > > it annoys me for a long time, so let me ask: > > Does anybody have a tip on a repo with a distribution that is small? > > Unfortunately Pulp does not provide to create a distribution. The only > way to get it is to sync something. I am using Fedora for my testing. > > I did modify Tomas' ZOO repo and added some package groups, categories > and files (empty iso file). But I was not able to find a simple way how > to create a small repo with a distribution. > > http://lzap.fedorapeople.org/fakerepos/zoo2/ > > I'd be happy for any hints. I would love to include also distributions > in our cli "smoke" tests, but it is not possible to sync Fedora each > run. > From tstrachota at redhat.com Mon Dec 5 15:01:03 2011 From: tstrachota at redhat.com (Tomas Strachota) Date: Mon, 05 Dec 2011 16:01:03 +0100 Subject: [katello-devel] Small repo with distribution In-Reply-To: <4EDCDA7E.2050209@redhat.com> References: <20111205144030.GF26396@lzapx.brq.redhat.com> <4EDCDA7E.2050209@redhat.com> Message-ID: <4EDCDCAF.2040505@redhat.com> On 12/05/2011 03:51 PM, Shannon Hughes wrote: > i used some of John Matthews small repos when testing sync error > conditions. He might still have them up or can assist. > > On 12/05/2011 09:40 AM, Lukas Zapletal wrote: >> Hey, >> >> it annoys me for a long time, so let me ask: >> >> Does anybody have a tip on a repo with a distribution that is small? >> >> Unfortunately Pulp does not provide to create a distribution. The only >> way to get it is to sync something. I am using Fedora for my testing. >> I also use fedora repos and synchronize ONLY distributions via pulp's api. You can restrict what to synchronize in the parameter 'skip'. It's much faster. Eg. {"skip": {"packages": 1, "errata": 1}} See: https://fedorahosted.org/pulp/wiki/UGREST-Repositories#RepositorySychronization1 >> I did modify Tomas' ZOO repo and added some package groups, categories >> and files (empty iso file). But I was not able to find a simple way how >> to create a small repo with a distribution. >> >> http://lzap.fedorapeople.org/fakerepos/zoo2/ >> >> I'd be happy for any hints. I would love to include also distributions >> in our cli "smoke" tests, but it is not possible to sync Fedora each >> run. >> > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel From jmatthew at redhat.com Mon Dec 5 15:14:38 2011 From: jmatthew at redhat.com (John Matthews) Date: Mon, 05 Dec 2011 10:14:38 -0500 (EST) Subject: [katello-devel] Small repo with distribution In-Reply-To: <4EDCDA7E.2050209@redhat.com> Message-ID: You can fake out a distribution by creating a .treeinfo file and manually adding checksum lines corresponding to files you want in the distribution. Here is an example of what I use when testing: http://jmatthews.fedorapeople.org/pulp_unittest/.treeinfo sudo pulp-admin repo create --id pulp_unittest --feed http://jmatthews.fedorapeople.org/pulp_unittest/ ----- Original Message ----- > i used some of John Matthews small repos when testing sync error > conditions. He might still have them up or can assist. > > On 12/05/2011 09:40 AM, Lukas Zapletal wrote: > > Hey, > > > > it annoys me for a long time, so let me ask: > > > > Does anybody have a tip on a repo with a distribution that is > > small? > > > > Unfortunately Pulp does not provide to create a distribution. The > > only > > way to get it is to sync something. I am using Fedora for my > > testing. > > > > I did modify Tomas' ZOO repo and added some package groups, > > categories > > and files (empty iso file). But I was not able to find a simple way > > how > > to create a small repo with a distribution. > > > > http://lzap.fedorapeople.org/fakerepos/zoo2/ > > > > I'd be happy for any hints. I would love to include also > > distributions > > in our cli "smoke" tests, but it is not possible to sync Fedora > > each > > run. > > > > From pkilambi at redhat.com Mon Dec 5 15:24:52 2011 From: pkilambi at redhat.com (Pradeep Kilambi) Date: Mon, 05 Dec 2011 10:24:52 -0500 Subject: [katello-devel] Small repo with distribution In-Reply-To: <4EDCDCAF.2040505@redhat.com> References: <20111205144030.GF26396@lzapx.brq.redhat.com> <4EDCDA7E.2050209@redhat.com> <4EDCDCAF.2040505@redhat.com> Message-ID: <4EDCE244.2090407@redhat.com> On 12/05/2011 10:01 AM, Tomas Strachota wrote: > On 12/05/2011 03:51 PM, Shannon Hughes wrote: >> i used some of John Matthews small repos when testing sync error >> conditions. He might still have them up or can assist. >> >> On 12/05/2011 09:40 AM, Lukas Zapletal wrote: >>> Hey, >>> >>> it annoys me for a long time, so let me ask: >>> >>> Does anybody have a tip on a repo with a distribution that is small? >>> >>> Unfortunately Pulp does not provide to create a distribution. The only >>> way to get it is to sync something. I am using Fedora for my testing. >>> > > I also use fedora repos and synchronize ONLY distributions via pulp's > api. You can restrict what to synchronize in the parameter 'skip'. > It's much faster. > Eg. {"skip": {"packages": 1, "errata": 1}} > > See: > https://fedorahosted.org/pulp/wiki/UGREST-Repositories#RepositorySychronization1 > Thats correct. The best way is to limit what you wanna sync by skipping packages and errata and allow only distributions. Also another way is, if you already have a repo synced in pulp, create a feedless repo, add distribution to it and then export it via pulp and just expose this repo publicly for further use. Lemme know if you need further help. ~ Prad > >>> I did modify Tomas' ZOO repo and added some package groups, categories >>> and files (empty iso file). But I was not able to find a simple way how >>> to create a small repo with a distribution. >>> >>> http://lzap.fedorapeople.org/fakerepos/zoo2/ >>> >>> I'd be happy for any hints. I would love to include also distributions >>> in our cli "smoke" tests, but it is not possible to sync Fedora each >>> run. >>> >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel From lzap at redhat.com Mon Dec 5 17:18:18 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Mon, 5 Dec 2011 18:18:18 +0100 Subject: [katello-devel] Small repo with distribution In-Reply-To: References: <4EDCDA7E.2050209@redhat.com> Message-ID: <20111205171818.GI26396@lzapx.brq.redhat.com> John, this is cool. I have created a new zoo3 repo that contains a distribution now :-) http://lzap.fedorapeople.org/fakerepos/zoo3/ I did not touch the timestamp, I hope it works. http://lzap.fedorapeople.org/fakerepos/zoo3/.treeinfo I will incorporate it into our tests, thanks. LZ On Mon, Dec 05, 2011 at 10:14:38AM -0500, John Matthews wrote: > You can fake out a distribution by creating a .treeinfo file and manually adding checksum lines corresponding to files you want in the distribution. > > Here is an example of what I use when testing: > http://jmatthews.fedorapeople.org/pulp_unittest/.treeinfo > > sudo pulp-admin repo create --id pulp_unittest --feed http://jmatthews.fedorapeople.org/pulp_unittest/ > > > ----- Original Message ----- > > i used some of John Matthews small repos when testing sync error > > conditions. He might still have them up or can assist. > > > > On 12/05/2011 09:40 AM, Lukas Zapletal wrote: > > > Hey, > > > > > > it annoys me for a long time, so let me ask: > > > > > > Does anybody have a tip on a repo with a distribution that is > > > small? > > > > > > Unfortunately Pulp does not provide to create a distribution. The > > > only > > > way to get it is to sync something. I am using Fedora for my > > > testing. > > > > > > I did modify Tomas' ZOO repo and added some package groups, > > > categories > > > and files (empty iso file). But I was not able to find a simple way > > > how > > > to create a small repo with a distribution. > > > > > > http://lzap.fedorapeople.org/fakerepos/zoo2/ > > > > > > I'd be happy for any hints. I would love to include also > > > distributions > > > in our cli "smoke" tests, but it is not possible to sync Fedora > > > each > > > run. > > > > > > > > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From bkearney at redhat.com Mon Dec 5 17:25:22 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Mon, 05 Dec 2011 12:25:22 -0500 Subject: [katello-devel] Fully qualified domain name In-Reply-To: <20111202170830.GI3395@lzapx.brq.redhat.com> References: <20111202170830.GI3395@lzapx.brq.redhat.com> Message-ID: <4EDCFE82.501@redhat.com> On 12/02/2011 12:08 PM, Lukas Zapletal wrote: > Hello, > > it seems that the seed error we see very often in Puppet occurs, when > domainname is not set properly. Before running katello-configure please > check the following things: > > 1) You have FQDN set > > $ hostname > myhostname.mydomain > > 2) You have both in the /etc/hosts or your DNS resolves everything > properly: > > $ ping -c1 localhost > $ ping -c1 myhostname.mydomain > > I will add a check to the katello-configure script that will throw an > error. More info at: > > https://bugzilla.redhat.com/show_bug.cgi?id=758712 > > Thank to gszasz and inecas for helping with tracking this ugly one. > Do the Headpin/Katello install docs need to be updated? -- bk From mmccune at redhat.com Mon Dec 5 18:39:37 2011 From: mmccune at redhat.com (Mike McCune) Date: Mon, 05 Dec 2011 10:39:37 -0800 Subject: [katello-devel] katello-configure currently blocked Message-ID: <4EDD0FE9.5000403@redhat.com> FYI if anyone tries to do an install today they will most likely run into this error installing: # katello-configure Starting Katello configuration The top-level log file is [/var/log/katello/katello-configure-20111205-130038/main.log] err: /Stage[main]/Certs::Config/Exec[generate-ssl-keystore]/returns: change from notrun to 0 failed: openssl pkcs12 -export -in /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -inkey /root/ssl-build/KATELLO-PRIVATE-SSL-KEY -out /etc/pki/katello/keystore -name tomcat -CAfile /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -caname root -chain -passin "file:/etc/katello/ssl_ca_password-file" -password "file:/etc/katello/keystore_password-file" returned 1 instead of one of [0] at /usr/share/katello/install/puppet/modules/certs/manifests/config.pp:102 documented here: https://bugzilla.redhat.com/show_bug.cgi?id=760280 looking into what is going on and will report back when things have cleared up. Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From mmccune at redhat.com Mon Dec 5 19:28:20 2011 From: mmccune at redhat.com (Mike McCune) Date: Mon, 05 Dec 2011 11:28:20 -0800 Subject: [katello-devel] katello-configure currently blocked In-Reply-To: <4EDD0FE9.5000403@redhat.com> References: <4EDD0FE9.5000403@redhat.com> Message-ID: <4EDD1B54.3070908@redhat.com> On 12/05/2011 10:39 AM, Mike McCune wrote: > FYI if anyone tries to do an install today they will most likely run > into this error installing: > > # katello-configure > Starting Katello configuration > The top-level log file is > [/var/log/katello/katello-configure-20111205-130038/main.log] > err: /Stage[main]/Certs::Config/Exec[generate-ssl-keystore]/returns: change > from notrun to 0 failed: openssl pkcs12 -export -in > /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -inkey > /root/ssl-build/KATELLO-PRIVATE-SSL-KEY -out /etc/pki/katello/keystore -name > tomcat -CAfile /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -caname root > -chain > -passin "file:/etc/katello/ssl_ca_password-file" -password > "file:/etc/katello/keystore_password-file" returned 1 instead of one of > [0] at > /usr/share/katello/install/puppet/modules/certs/manifests/config.pp:102 > > documented here: > > https://bugzilla.redhat.com/show_bug.cgi?id=760280 > > looking into what is going on and will report back when things have > cleared up. > > Mike quick workaround : # katello-configure --skip-ssl-ca-generation=True Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From jweiss at redhat.com Mon Dec 5 19:40:43 2011 From: jweiss at redhat.com (Jeff Weiss) Date: Mon, 05 Dec 2011 14:40:43 -0500 Subject: [katello-devel] katello-configure currently blocked In-Reply-To: <4EDD1B54.3070908@redhat.com> References: <4EDD0FE9.5000403@redhat.com> <4EDD1B54.3070908@redhat.com> Message-ID: <1323114043.17548.37.camel@jweiss> On Mon, 2011-12-05 at 11:28 -0800, Mike McCune wrote: > On 12/05/2011 10:39 AM, Mike McCune wrote: > > FYI if anyone tries to do an install today they will most likely run > > into this error installing: > > > > # katello-configure > > Starting Katello configuration > > The top-level log file is > > [/var/log/katello/katello-configure-20111205-130038/main.log] > > err: /Stage[main]/Certs::Config/Exec[generate-ssl-keystore]/returns: change > > from notrun to 0 failed: openssl pkcs12 -export -in > > /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -inkey > > /root/ssl-build/KATELLO-PRIVATE-SSL-KEY -out /etc/pki/katello/keystore -name > > tomcat -CAfile /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -caname root > > -chain > > -passin "file:/etc/katello/ssl_ca_password-file" -password > > "file:/etc/katello/keystore_password-file" returned 1 instead of one of > > [0] at > > /usr/share/katello/install/puppet/modules/certs/manifests/config.pp:102 > > > > documented here: > > > > https://bugzilla.redhat.com/show_bug.cgi?id=760280 > > > > looking into what is going on and will report back when things have > > cleared up. > > > > Mike > > quick workaround : > > # katello-configure --skip-ssl-ca-generation=True > > Mike FYI, the CI install of 08ec9ada86f6f25fc66164b625d878babac239a1 worked: http://hudson.rhq.lab.eng.bos.redhat.com:8080/hudson/job/katello-ci-install/249/console -jeff From cperry at redhat.com Mon Dec 5 20:49:43 2011 From: cperry at redhat.com (Cliff Perry) Date: Mon, 05 Dec 2011 15:49:43 -0500 Subject: [katello-devel] katello-configure currently blocked In-Reply-To: <1323114043.17548.37.camel@jweiss> References: <4EDD0FE9.5000403@redhat.com> <4EDD1B54.3070908@redhat.com> <1323114043.17548.37.camel@jweiss> Message-ID: <4EDD2E67.3030105@redhat.com> On 12/05/2011 02:40 PM, Jeff Weiss wrote: > On Mon, 2011-12-05 at 11:28 -0800, Mike McCune wrote: >> On 12/05/2011 10:39 AM, Mike McCune wrote: >>> FYI if anyone tries to do an install today they will most likely run >>> into this error installing: >>> >>> # katello-configure >>> Starting Katello configuration >>> The top-level log file is >>> [/var/log/katello/katello-configure-20111205-130038/main.log] >>> err: /Stage[main]/Certs::Config/Exec[generate-ssl-keystore]/returns: change >>> from notrun to 0 failed: openssl pkcs12 -export -in >>> /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -inkey >>> /root/ssl-build/KATELLO-PRIVATE-SSL-KEY -out /etc/pki/katello/keystore -name >>> tomcat -CAfile /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -caname root >>> -chain >>> -passin "file:/etc/katello/ssl_ca_password-file" -password >>> "file:/etc/katello/keystore_password-file" returned 1 instead of one of >>> [0] at >>> /usr/share/katello/install/puppet/modules/certs/manifests/config.pp:102 >>> >>> documented here: >>> >>> https://bugzilla.redhat.com/show_bug.cgi?id=760280 >>> >>> looking into what is going on and will report back when things have >>> cleared up. >>> >>> Mike >> >> quick workaround : >> >> # katello-configure --skip-ssl-ca-generation=True >> >> Mike > > FYI, the CI install of 08ec9ada86f6f25fc66164b625d878babac239a1 worked: > http://hudson.rhq.lab.eng.bos.redhat.com:8080/hudson/job/katello-ci-install/249/console > > -jeff > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel My test install completed without error. [root at dhcp77-206 ~]# katello-configure --help Usage: /usr/sbin/katello-configure [options] --answer-file=ANSWER_FILE Path of the answer file --user-name=USER_NAME Katello administrative user (default: admin) --user-pass=USER_PASS Katello user's password (default: admin) --user-email=USER_EMAIL Katello user's email (default: root at localhost) --org-name=ORG_NAME Katello initial Organization (default: ACME_Corporation) --proxy-url=PROXY_URL HTTP Proxy URL (example: http://172.31.1.1) --proxy-port=PROXY_PORT HTTP Proxy port (default: 3128) --proxy-user=PROXY_USER HTTP Proxy user (proxy username, if auth is required) --proxy-pass=PROXY_PASS HTTP Proxy pass (proxy password, if auth is required) --db-name=DB_NAME Katello database name --db-user=DB_USER Katello database user --db-password=DB_PASSWORD Katello database password --deployment=DEPLOYMENT Deployment Type --non-interactive=NON_INTERACTIVE Non-interactive installer mode --skip-ssl-ca-generation=SKIP_SSL_CA_GENERATION skip SSL CA generation --ssl-ca-password=SSL_CA_PASSWORD SSL CA password --ssl-ca-country=SSL_CA_COUNTRY SSL CA country --ssl-ca-state=SSL_CA_STATE SSL CA state --ssl-ca-city=SSL_CA_CITY SSL CA city --ssl-ca-org=SSL_CA_ORG SSL CA organization --ssl-ca-org-unit=SSL_CA_ORG_UNIT SSL CA organization unit --ssl-ca-cn=SSL_CA_CN SSL CA common name --ssl-ca-email=SSL_CA_EMAIL SSL CA e-mail address --ssl-cert-expiration=SSL_CERT_EXPIRATION SSL certificate expiration (in days) --ssl-ca-password-file=SSL_CA_PASSWORD_FILE SSL CA password file path --keystore-password-file=KEYSTORE_PASSWORD_FILE Keystore password file path --nss-db-password-file=NSS_DB_PASSWORD_FILE NSS DB password file path --only-show-config Print the resulting configuration and exit -h, --help Show this short summary [root at dhcp77-206 ~]# hostname dhcp77-206.rhndev.redhat.com [root at dhcp77-206 ~]# katello-configure Starting Katello configuration The top-level log file is [/var/log/katello/katello-configure-20111205-152404/main.log] [root at dhcp77-206 ~]# echo $? 0 [root at dhcp77-206 ~]# rpm -q katello-configure katello-configure-0.1.24-1.el6.noarch [root at dhcp77-206 ~]# From lzap at redhat.com Tue Dec 6 13:56:29 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 6 Dec 2011 14:56:29 +0100 Subject: [katello-devel] Fully qualified domain name In-Reply-To: <4EDCFE82.501@redhat.com> References: <20111202170830.GI3395@lzapx.brq.redhat.com> <4EDCFE82.501@redhat.com> Message-ID: <20111206135629.GB3471@lzapx.brq.redhat.com> > Do the Headpin/Katello install docs need to be updated? Yes, let me do it now: https://fedorahosted.org/katello/wiki/Install?action=diff&version=98 There is one more issue - when FQDN is not set, Puppet "determines" the domain from the search entry in the /etc/resolv.conf for some reason. This is pretty dangerous and unexpected behaviour. I guess we need to "override" facter variables: domainname and fqdn. https://bugzilla.redhat.com/show_bug.cgi?id=760265 -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From jweiss at redhat.com Tue Dec 6 15:43:00 2011 From: jweiss at redhat.com (Jeff Weiss) Date: Tue, 06 Dec 2011 10:43:00 -0500 Subject: [katello-devel] changeset API breakage Message-ID: <1323186180.10548.5.camel@blinky.rdu.redhat.com> My automation uses API to create changesets and promote them- which suddenly broke after Tomas's merge http://hudson.rhq.lab.eng.bos.redhat.com:8080/hudson/job/katello-unit/2063/changes The error I get now is: Started PUT "/katello//api/changesets/1" for 10.16.120.31 at Tue Dec 06 09:00:41 -0500 2011 Processing by ErrorsController#routing as JSON Parameters: {"a"=>"api/changesets/1", "patch"=>{"+repos"=>nil, "+errata"=>nil, "+templates"=>nil, "+packages"=>nil, "+products"=>["templateProduct-1323180038641", "templateProduct-1323180038642", "templateProduct-1323180038643"]}} Completed 404 Not Found in 2ms Is this a bug or did the API change? Jeff -- Jeff Weiss Principal Quality Assurance Engineer jweiss at redhat.com (919)754-4178 From lzap at redhat.com Tue Dec 6 16:14:17 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 6 Dec 2011 17:14:17 +0100 Subject: [katello-devel] Katello SSL issues Message-ID: <20111206161417.GC3471@lzapx.brq.redhat.com> Hey, as I wrote yesterday, I accepted Tomas' patch. We have now the tool and few Puppet steps which generate CA cert and Apache http/Qpid certs, but they are unused right now. I would recommend to postpone work on SSL certs until end of next week. Then we can re-enable them and start with tuning and testing. I expect some changes on the Candlepin/Pulp side, because the plan is to have "the big" Katello CA and have all certificates signed with it. That could possibly cause troubles (instead checking one CA clients will need to check the whole chain). Instead of big bang, I would stick with step-by-step approach. We can start with Pulp as I think the integration should be easier and then move on with Candlepin (that's gonna hurt). By the way there is/was an issue with our SSL generation due to race condition/working directory issue in Puppet. I hope this is solved now, but until packages get built, please make sure you start katello-configure in the /root directory: pushd /root && katello-configure && popd https://bugzilla.redhat.com/show_bug.cgi?id=760280 Please not there is also another bug, if you do not have FQDN properly set, Puppet "guess" it from /etc/resolv.conf which cause troubles. Fix have been commited moments ago: https://bugzilla.redhat.com/show_bug.cgi?id=760265 In short - katello-configure-0.1.26-1 should be fine. Please report here. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From tstrachota at redhat.com Tue Dec 6 17:49:13 2011 From: tstrachota at redhat.com (Tomas Strachota) Date: Tue, 06 Dec 2011 18:49:13 +0100 Subject: [katello-devel] changeset API breakage In-Reply-To: <1323186180.10548.5.camel@blinky.rdu.redhat.com> References: <1323186180.10548.5.camel@blinky.rdu.redhat.com> Message-ID: <4EDE5599.3030408@redhat.com> On 12/06/2011 04:43 PM, Jeff Weiss wrote: > My automation uses API to create changesets and promote them- which > suddenly broke after Tomas's merge > > http://hudson.rhq.lab.eng.bos.redhat.com:8080/hudson/job/katello-unit/2063/changes > > The error I get now is: > > Started PUT "/katello//api/changesets/1" for 10.16.120.31 at Tue Dec 06 > 09:00:41 -0500 2011 > Processing by ErrorsController#routing as JSON > Parameters: {"a"=>"api/changesets/1", "patch"=>{"+repos"=>nil, > "+errata"=>nil, "+templates"=>nil, "+packages"=>nil, > "+products"=>["templateProduct-1323180038641", > "templateProduct-1323180038642", "templateProduct-1323180038643"]}} > Completed 404 Not Found in 2ms > > Is this a bug or did the API change? > > Jeff > It's API change. Sorry for not announcing earlier! The api has been changed to use more RESTy way of updating content. Similar to templates, there's new changesets_content controller. I added new routes to this brief API guide [1]. I'll try to put more complete documentation somewhere as this list of routes does not say much without browsing the controller code. Tomas [1] https://fedorahosted.org/katello/wiki/APIGuide From jweiss at redhat.com Tue Dec 6 21:29:57 2011 From: jweiss at redhat.com (Jeff Weiss) Date: Tue, 06 Dec 2011 16:29:57 -0500 Subject: [katello-devel] Katello SSL issues In-Reply-To: <20111206161417.GC3471@lzapx.brq.redhat.com> References: <20111206161417.GC3471@lzapx.brq.redhat.com> Message-ID: <1323206997.10548.13.camel@blinky.rdu.redhat.com> On Tue, 2011-12-06 at 17:14 +0100, Lukas Zapletal wrote: > Hey, > > as I wrote yesterday, I accepted Tomas' patch. We have now the tool and > few Puppet steps which generate CA cert and Apache http/Qpid certs, but > they are unused right now. > > I would recommend to postpone work on SSL certs until end of next week. > Then we can re-enable them and start with tuning and testing. I expect > some changes on the Candlepin/Pulp side, because the plan is to have > "the big" Katello CA and have all certificates signed with it. That > could possibly cause troubles (instead checking one CA clients will need > to check the whole chain). Instead of big bang, I would stick with > step-by-step approach. We can start with Pulp as I think the integration > should be easier and then move on with Candlepin (that's gonna hurt). > > By the way there is/was an issue with our SSL generation due to race > condition/working directory issue in Puppet. I hope this is solved now, > but until packages get built, please make sure you start > katello-configure in the /root directory: > > pushd /root && katello-configure && popd > > https://bugzilla.redhat.com/show_bug.cgi?id=760280 > > Please not there is also another bug, if you do not have FQDN properly > set, Puppet "guess" it from /etc/resolv.conf which cause troubles. Fix > have been commited moments ago: > > https://bugzilla.redhat.com/show_bug.cgi?id=760265 > > In short - katello-configure-0.1.26-1 should be fine. Please report > here. > I've done a lot of testing with SSL already - browsers don't like self-signed certs, and automated testing with Selenium can't get past the warning by itself. So I started using my own CA to sign the candlepin certs. With a bit of scripting to sign and copy the certs to the right places, the Katello UI works fine. With a bit more configuring of apache, RHSM clients can register. But I got totally stuck where Pulp cannot validate the client certs coming from RHSM. I believe the problem may be that pulp will only check against one cert, not the whole chain (it fails if there's more than 1 in the chain). The katello installer sets up pulp to point at Candlepin's cert but doesn't include my CA in the chain. But it's more than that, I suspect. I don't think pulp is even capable of pointing at multiple CA files, and it doesn't seem to accept them concatenated together the way most SSL libs do. Jay Dobies was helping me out with this a few weeks ago but we did not finish getting to the root of the problem. -- Jeff Weiss Principal Quality Assurance Engineer jweiss at redhat.com (919)754-4178 From lzap+pub at redhat.com Tue Dec 6 22:35:55 2011 From: lzap+pub at redhat.com (Lukas Zapletal) Date: Tue, 6 Dec 2011 23:35:55 +0100 Subject: [katello-devel] katello-configure currently blocked In-Reply-To: <1323114043.17548.37.camel@jweiss> References: <4EDD0FE9.5000403@redhat.com> <4EDD1B54.3070908@redhat.com> <1323114043.17548.37.camel@jweiss> Message-ID: <20111206223555.GE4789@lzapx.brq.redhat.com> Yup, katello-configure-0.1.26-1 fixes this problem FYI. LZ On Mon, Dec 05, 2011 at 02:40:43PM -0500, Jeff Weiss wrote: > On Mon, 2011-12-05 at 11:28 -0800, Mike McCune wrote: > > On 12/05/2011 10:39 AM, Mike McCune wrote: > > > FYI if anyone tries to do an install today they will most likely run > > > into this error installing: > > > > > > # katello-configure > > > Starting Katello configuration > > > The top-level log file is > > > [/var/log/katello/katello-configure-20111205-130038/main.log] > > > err: /Stage[main]/Certs::Config/Exec[generate-ssl-keystore]/returns: change > > > from notrun to 0 failed: openssl pkcs12 -export -in > > > /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -inkey > > > /root/ssl-build/KATELLO-PRIVATE-SSL-KEY -out /etc/pki/katello/keystore -name > > > tomcat -CAfile /usr/share/katello/KATELLO-TRUSTED-SSL-CERT -caname root > > > -chain > > > -passin "file:/etc/katello/ssl_ca_password-file" -password > > > "file:/etc/katello/keystore_password-file" returned 1 instead of one of > > > [0] at > > > /usr/share/katello/install/puppet/modules/certs/manifests/config.pp:102 > > > > > > documented here: > > > > > > https://bugzilla.redhat.com/show_bug.cgi?id=760280 > > > > > > looking into what is going on and will report back when things have > > > cleared up. > > > > > > Mike > > > > quick workaround : > > > > # katello-configure --skip-ssl-ca-generation=True > > > > Mike > > FYI, the CI install of 08ec9ada86f6f25fc66164b625d878babac239a1 worked: > http://hudson.rhq.lab.eng.bos.redhat.com:8080/hudson/job/katello-ci-install/249/console > > -jeff > > _______________________________________________ > katello mailing list > katello at lists.fedorahosted.org > https://fedorahosted.org/mailman/listinfo/katello -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From mmccune at redhat.com Wed Dec 7 02:08:01 2011 From: mmccune at redhat.com (Mike McCune) Date: Tue, 06 Dec 2011 18:08:01 -0800 Subject: [katello-devel] pulp-testing needed if you want to promote distributions Message-ID: <4EDECA81.1050701@redhat.com> Currently Katello requires an updated API path for pulp's distributions: diff --git a/src/lib/resources/pulp.rb b/src/lib/resources/pulp.rb index 314d2a0..568c2dd 100644 --- a/src/lib/resources/pulp.rb +++ b/src/lib/resources/pulp.rb @@ -145,7 +145,7 @@ module Pulp end def dist_path - "/pulp/api/distributions/" + "/pulp/api/distribution/" end end end the current community release of Pulp still points at /distribution/ so in order to promote distributions you will need to switch to the pulp-testing repo as found in your: /etc/yum.repos.d/pulp.repo: # Testing Builds (not necessarily nightly but more frequently than a full blown # community release) [pulp-testing] name=Pulp Testing Builds baseurl=http://repos.fedorapeople.org/repos/pulp/pulp/testing/$releasever/$basearch/ enabled=1 skip_if_unavailable=1 gpgcheck=0 Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From lzap at redhat.com Wed Dec 7 09:35:40 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Wed, 7 Dec 2011 10:35:40 +0100 Subject: [katello-devel] Katello SSL issues In-Reply-To: <1323206997.10548.13.camel@blinky.rdu.redhat.com> References: <20111206161417.GC3471@lzapx.brq.redhat.com> <1323206997.10548.13.camel@blinky.rdu.redhat.com> Message-ID: <20111207093540.GB19321@lzapx.brq.redhat.com> Jeff, those are good points. It looks like a sprint task for Pulp team to get it working. Now, correct me if I am wrong, but our main idea is to have one "main" (Katello) CA that will sign all the others. LZ On Tue, Dec 06, 2011 at 04:29:57PM -0500, Jeff Weiss wrote: > On Tue, 2011-12-06 at 17:14 +0100, Lukas Zapletal wrote: > > Hey, > > > > as I wrote yesterday, I accepted Tomas' patch. We have now the tool and > > few Puppet steps which generate CA cert and Apache http/Qpid certs, but > > they are unused right now. > > > > I would recommend to postpone work on SSL certs until end of next week. > > Then we can re-enable them and start with tuning and testing. I expect > > some changes on the Candlepin/Pulp side, because the plan is to have > > "the big" Katello CA and have all certificates signed with it. That > > could possibly cause troubles (instead checking one CA clients will need > > to check the whole chain). Instead of big bang, I would stick with > > step-by-step approach. We can start with Pulp as I think the integration > > should be easier and then move on with Candlepin (that's gonna hurt). > > > > By the way there is/was an issue with our SSL generation due to race > > condition/working directory issue in Puppet. I hope this is solved now, > > but until packages get built, please make sure you start > > katello-configure in the /root directory: > > > > pushd /root && katello-configure && popd > > > > https://bugzilla.redhat.com/show_bug.cgi?id=760280 > > > > Please not there is also another bug, if you do not have FQDN properly > > set, Puppet "guess" it from /etc/resolv.conf which cause troubles. Fix > > have been commited moments ago: > > > > https://bugzilla.redhat.com/show_bug.cgi?id=760265 > > > > In short - katello-configure-0.1.26-1 should be fine. Please report > > here. > > > > I've done a lot of testing with SSL already - browsers don't like > self-signed certs, and automated testing with Selenium can't get past > the warning by itself. So I started using my own CA to sign the > candlepin certs. > > With a bit of scripting to sign and copy the certs to the right places, > the Katello UI works fine. With a bit more configuring of apache, RHSM > clients can register. > > But I got totally stuck where Pulp cannot validate the client certs > coming from RHSM. I believe the problem may be that pulp will only > check against one cert, not the whole chain (it fails if there's more > than 1 in the chain). The katello installer sets up pulp to point at > Candlepin's cert but doesn't include my CA in the chain. But it's more > than that, I suspect. I don't think pulp is even capable of pointing at > multiple CA files, and it doesn't seem to accept them concatenated > together the way most SSL libs do. > > Jay Dobies was helping me out with this a few weeks ago but we did not > finish getting to the root of the problem. > > > -- > Jeff Weiss > Principal Quality Assurance Engineer > jweiss at redhat.com > (919)754-4178 > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Wed Dec 7 12:00:51 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Wed, 7 Dec 2011 13:00:51 +0100 Subject: [katello-devel] Stealing awesome git-untagged-commits.pl Message-ID: <20111207120051.GD19321@lzapx.brq.redhat.com> Guys, I am stealing the rel-eng/git-untagged-commits.pl script from Spacewalk repo and putting it in the same place in Katello. What it does? It shows all untagged commits. As you already noticed, we started building from tags only. No tag = no build = no fun. Example: $ rel-eng/git-untagged-commits.pl katello-cli-tests-0.1.10-1..HEAD: 8c68b66 cli tests - switching to zoo3 test repo (lzap+git at redhat.com) ------------------------------------------ Here we go! That guy named lzap forgot to tag katello-cli-tests! Shame on him. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Wed Dec 7 12:17:14 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Wed, 7 Dec 2011 13:17:14 +0100 Subject: [katello-devel] Small repo with distribution In-Reply-To: <20111205171818.GI26396@lzapx.brq.redhat.com> References: <4EDCDA7E.2050209@redhat.com> <20111205171818.GI26396@lzapx.brq.redhat.com> Message-ID: <20111207121714.GE19321@lzapx.brq.redhat.com> It does not work for me. Is the issue in having files out of images/ directory? LZ On Mon, Dec 05, 2011 at 06:18:18PM +0100, Lukas Zapletal wrote: > John, this is cool. > > I have created a new zoo3 repo that contains a distribution now :-) > > http://lzap.fedorapeople.org/fakerepos/zoo3/ > > I did not touch the timestamp, I hope it works. > > http://lzap.fedorapeople.org/fakerepos/zoo3/.treeinfo > > I will incorporate it into our tests, thanks. > > LZ > > On Mon, Dec 05, 2011 at 10:14:38AM -0500, John Matthews wrote: > > You can fake out a distribution by creating a .treeinfo file and manually adding checksum lines corresponding to files you want in the distribution. > > > > Here is an example of what I use when testing: > > http://jmatthews.fedorapeople.org/pulp_unittest/.treeinfo > > > > sudo pulp-admin repo create --id pulp_unittest --feed http://jmatthews.fedorapeople.org/pulp_unittest/ > > > > > > ----- Original Message ----- > > > i used some of John Matthews small repos when testing sync error > > > conditions. He might still have them up or can assist. > > > > > > On 12/05/2011 09:40 AM, Lukas Zapletal wrote: > > > > Hey, > > > > > > > > it annoys me for a long time, so let me ask: > > > > > > > > Does anybody have a tip on a repo with a distribution that is > > > > small? > > > > > > > > Unfortunately Pulp does not provide to create a distribution. The > > > > only > > > > way to get it is to sync something. I am using Fedora for my > > > > testing. > > > > > > > > I did modify Tomas' ZOO repo and added some package groups, > > > > categories > > > > and files (empty iso file). But I was not able to find a simple way > > > > how > > > > to create a small repo with a distribution. > > > > > > > > http://lzap.fedorapeople.org/fakerepos/zoo2/ > > > > > > > > I'd be happy for any hints. I would love to include also > > > > distributions > > > > in our cli "smoke" tests, but it is not possible to sync Fedora > > > > each > > > > run. > > > > > > > > > > > > > > _______________________________________________ > > katello-devel mailing list > > katello-devel at redhat.com > > https://www.redhat.com/mailman/listinfo/katello-devel > > -- > Later, > > Lukas Zapletal | E32E400A > RHN Satellite Engineering > Red Hat Czech s.r.o. Brno > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Wed Dec 7 13:44:12 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Wed, 7 Dec 2011 14:44:12 +0100 Subject: [katello-devel] Small repo with distribution In-Reply-To: <20111207121714.GE19321@lzapx.brq.redhat.com> References: <4EDCDA7E.2050209@redhat.com> <20111205171818.GI26396@lzapx.brq.redhat.com> <20111207121714.GE19321@lzapx.brq.redhat.com> Message-ID: <20111207134412.GF19321@lzapx.brq.redhat.com> Thanks John for your help on the channel. I pushed a fixed small repo here: http://lzap.fedorapeople.org/fakerepos/zoo4/ It contains: - packages - groups - categories - one errata - few files - one distribution LZ On Wed, Dec 07, 2011 at 01:17:14PM +0100, Lukas Zapletal wrote: > It does not work for me. Is the issue in having files out of images/ > directory? > > LZ > > On Mon, Dec 05, 2011 at 06:18:18PM +0100, Lukas Zapletal wrote: > > John, this is cool. > > > > I have created a new zoo3 repo that contains a distribution now :-) > > > > http://lzap.fedorapeople.org/fakerepos/zoo3/ > > > > I did not touch the timestamp, I hope it works. > > > > http://lzap.fedorapeople.org/fakerepos/zoo3/.treeinfo > > > > I will incorporate it into our tests, thanks. > > > > LZ > > > > On Mon, Dec 05, 2011 at 10:14:38AM -0500, John Matthews wrote: > > > You can fake out a distribution by creating a .treeinfo file and manually adding checksum lines corresponding to files you want in the distribution. > > > > > > Here is an example of what I use when testing: > > > http://jmatthews.fedorapeople.org/pulp_unittest/.treeinfo > > > > > > sudo pulp-admin repo create --id pulp_unittest --feed http://jmatthews.fedorapeople.org/pulp_unittest/ > > > > > > > > > ----- Original Message ----- > > > > i used some of John Matthews small repos when testing sync error > > > > conditions. He might still have them up or can assist. > > > > > > > > On 12/05/2011 09:40 AM, Lukas Zapletal wrote: > > > > > Hey, > > > > > > > > > > it annoys me for a long time, so let me ask: > > > > > > > > > > Does anybody have a tip on a repo with a distribution that is > > > > > small? > > > > > > > > > > Unfortunately Pulp does not provide to create a distribution. The > > > > > only > > > > > way to get it is to sync something. I am using Fedora for my > > > > > testing. > > > > > > > > > > I did modify Tomas' ZOO repo and added some package groups, > > > > > categories > > > > > and files (empty iso file). But I was not able to find a simple way > > > > > how > > > > > to create a small repo with a distribution. > > > > > > > > > > http://lzap.fedorapeople.org/fakerepos/zoo2/ > > > > > > > > > > I'd be happy for any hints. I would love to include also > > > > > distributions > > > > > in our cli "smoke" tests, but it is not possible to sync Fedora > > > > > each > > > > > run. > > > > > > > > > > > > > > > > > > > _______________________________________________ > > > katello-devel mailing list > > > katello-devel at redhat.com > > > https://www.redhat.com/mailman/listinfo/katello-devel > > > > -- > > Later, > > > > Lukas Zapletal | E32E400A > > RHN Satellite Engineering > > Red Hat Czech s.r.o. Brno > > > > _______________________________________________ > > katello-devel mailing list > > katello-devel at redhat.com > > https://www.redhat.com/mailman/listinfo/katello-devel > > -- > Later, > > Lukas Zapletal | E32E400A > RHN Satellite Engineering > Red Hat Czech s.r.o. Brno > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Wed Dec 7 13:59:57 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Wed, 7 Dec 2011 14:59:57 +0100 Subject: [katello-devel] On changeset updating from CLI Message-ID: <20111207135957.GG19321@lzapx.brq.redhat.com> Hey, would it be a problem to have changeset names unique per organization? Right now we are unique per env per org. We have to type it in each update command in the CLI over and over again ;-) -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From jsherril at redhat.com Wed Dec 7 14:03:19 2011 From: jsherril at redhat.com (Justin Sherrill) Date: Wed, 07 Dec 2011 09:03:19 -0500 Subject: [katello-devel] On changeset updating from CLI In-Reply-To: <20111207135957.GG19321@lzapx.brq.redhat.com> References: <20111207135957.GG19321@lzapx.brq.redhat.com> Message-ID: <4EDF7227.5040703@redhat.com> On 12/07/2011 08:59 AM, Lukas Zapletal wrote: > Hey, > > would it be a problem to have changeset names unique per organization? > Right now we are unique per env per org. > > We have to type it in each update command in the CLI over and over again > ;-) > I would argue for unique per environment if we really want to make it unique at any level. -Justin From lzap at redhat.com Wed Dec 7 15:52:53 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Wed, 7 Dec 2011 16:52:53 +0100 Subject: [katello-devel] Pending system cli tests Message-ID: <20111207155253.GH19321@lzapx.brq.redhat.com> Backend, I started this wikipage https://fedorahosted.org/katello/wiki/PendingSystemTests to track the failing tests for us. Please update once you change the state. Fedora is green except one tests (Ivan is going to fix this hopefully). RHEL is also fine, but you need to update "yum" before run. The stock yum is not compatible with rhsm somehow. Documented. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From inecas at redhat.com Wed Dec 7 17:32:19 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Wed, 07 Dec 2011 18:32:19 +0100 Subject: [katello-devel] Import manifest orchestration reworked Message-ID: <4EDFA323.10701@redhat.com> Hi all, Few last days I was working on the rework of orchestration of manifest import and products promotion. Reasons why it needed to be changed and what the solution looks like are described on a wiki page [1]. In short, with this changes we don't need to change Candlepin content when promoting products/repos. Since this is quite a big change in our orchestration model, I would like you to review the changes and provide a feedback/comments to the current implementation. The changes are available in a remote branch *reimport-manifest*. [1] - https://fedorahosted.org/katello/wiki/CandlepinEnvironmentsIntegration PS: this changes also make our system tests green again:) -- Ivan -------------- next part -------------- An HTML attachment was scrubbed... URL: From jrist at redhat.com Wed Dec 7 17:40:15 2011 From: jrist at redhat.com (Jason Rist) Date: Wed, 07 Dec 2011 10:40:15 -0700 Subject: [katello-devel] Import manifest orchestration reworked In-Reply-To: <4EDFA323.10701@redhat.com> References: <4EDFA323.10701@redhat.com> Message-ID: <4EDFA4FF.6000604@redhat.com> On 12/07/2011 10:32 AM, Ivan Ne?as wrote: > Hi all, > > Few last days I was working on the rework of orchestration of manifest > import and products promotion. Reasons why it needed to be changed and > what the solution looks like are described on a wiki page [1]. In short, > with this changes we don't need to change Candlepin content when > promoting products/repos. > > Since this is quite a big change in our orchestration model, I would > like you to review the changes and provide a feedback/comments to the > current implementation. The changes are available in a remote branch > *reimport-manifest*. > > [1] - > https://fedorahosted.org/katello/wiki/CandlepinEnvironmentsIntegration > > PS: this changes also make our system tests green again:) > > -- Ivan > > > > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel Can we do a code review so everyone is on the same page with this? -J -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From jmatthew at redhat.com Wed Dec 7 20:11:59 2011 From: jmatthew at redhat.com (John Matthews) Date: Wed, 07 Dec 2011 15:11:59 -0500 (EST) Subject: [katello-devel] Avoid usage of '/tmp' In-Reply-To: <2d9ed619-cfc1-4286-b104-735185247c1f@zmail13.collab.prod.int.phx2.redhat.com> Message-ID: The issue of Grinder writing some data to /tmp/grinder and referring to it in between runs came up during our SELinux policy rewrite. Dan Walsh suggested we avoid using '/tmp/grinder' and instead switch to '/var/run/grinder'. I wanted to share his blog post highlighting this reason with the team. http://danwalsh.livejournal.com/11467.html " Daemon developers should follow these rules: /tmp is for users to store their stuff not for daemons or any process that is started in the boot process. If a daemon wants to communicate with a user then he should do it via /var/run/DAEMON. If you have a daemon that wants its temporarily files to survive a reboot. consider using /var/cache/DAEMON " Pulp's BZ to fix this: https://bugzilla.redhat.com/show_bug.cgi?id=761173 From lzap at redhat.com Thu Dec 8 09:22:23 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 8 Dec 2011 10:22:23 +0100 Subject: [katello-devel] Using Katello on systems without updates Message-ID: <20111208092223.GA8512@lzapx.brq.redhat.com> Hello, today I discovered few facts. If you try to install Katello on Fedora 15 with updates disabled, you get an dependency error (Puppet 2.6 required, but Fedora GOLD contains older version). On RHEL 6.1 with updates turned off there are two different issues. Although you are able to install, I am running into two issues: - if I restart pulp-sever mongodb segfaults which causes that lock file is blocking pulp from starting - after some decent amount of requests (our system test cli suite) tomcat falls down without any warning or error - the process just got killed With updates applied I don't have such issues. Instead of crying here, I am editing our wiki and putting a note for our users to update their systems before installing. https://fedorahosted.org/katello/wiki/Install I don't know why, but I am humming "Sign of Punk" song this morning ;-) http://www.youtube.com/watch?v=ZuYx_z9KU68 -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Thu Dec 8 11:44:21 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 8 Dec 2011 12:44:21 +0100 Subject: [katello-devel] Import manifest orchestration reworked In-Reply-To: <4EDFA323.10701@redhat.com> References: <4EDFA323.10701@redhat.com> Message-ID: <20111208114421.GC8512@lzapx.brq.redhat.com> So you also solved the issue when uebercert was blocking whole thing? No need of uebercert deletion anymore? >From my point of view, it looks very good. This change needs intensive testing it can have consequences we cannot currently think of. Bryan does it worth to include it in the drop? Because it's a good improvement. Good work! LZ On Wed, Dec 07, 2011 at 06:32:19PM +0100, Ivan Necas wrote: > Hi all, > > Few last days I was working on the rework of orchestration of > manifest import and products promotion. Reasons why it needed to be > changed and what the solution looks like are described on a wiki > page [1]. In short, with this changes we don't need to change > Candlepin content when promoting products/repos. > > Since this is quite a big change in our orchestration model, I would > like you to review the changes and provide a feedback/comments to > the current implementation. The changes are available in a remote > branch *reimport-manifest*. > > [1] - https://fedorahosted.org/katello/wiki/CandlepinEnvironmentsIntegration > > PS: this changes also make our system tests green again:) > > -- Ivan > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From bkearney at redhat.com Thu Dec 8 12:42:56 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Thu, 08 Dec 2011 07:42:56 -0500 Subject: [katello-devel] Using Katello on systems without updates In-Reply-To: <20111208092223.GA8512@lzapx.brq.redhat.com> References: <20111208092223.GA8512@lzapx.brq.redhat.com> Message-ID: <4EE0B0D0.3000509@redhat.com> On 12/08/2011 04:22 AM, Lukas Zapletal wrote: > Hello, > > today I discovered few facts. > > If you try to install Katello on Fedora 15 with updates disabled, you > get an dependency error (Puppet 2.6 required, but Fedora GOLD contains > older version). > > On RHEL 6.1 with updates turned off there are two different issues. > Although you are able to install, I am running into two issues: > > - if I restart pulp-sever mongodb segfaults which causes that lock file > is blocking pulp from starting > > - after some decent amount of requests (our system test cli suite) > tomcat falls down without any warning or error - the process just got > killed > > With updates applied I don't have such issues. > > Instead of crying here, I am editing our wiki and putting a note for our > users to update their systems before installing. I agree.. I think this is fine. > > https://fedorahosted.org/katello/wiki/Install I updated https://fedorahosted.org/candlepin/wiki/headpin/Install as well > > I don't know why, but I am humming "Sign of Punk" song this morning ;-) > > http://www.youtube.com/watch?v=ZuYx_z9KU68 > Too early. -- bk From lzap at redhat.com Thu Dec 8 13:13:11 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 8 Dec 2011 14:13:11 +0100 Subject: [katello-devel] Using Katello on systems without updates In-Reply-To: <4EE0B0D0.3000509@redhat.com> References: <20111208092223.GA8512@lzapx.brq.redhat.com> <4EE0B0D0.3000509@redhat.com> Message-ID: <20111208131311.GE8512@lzapx.brq.redhat.com> One more information - my Fedora issue (bad puppet version) is valid, but it turns out I had my swap off on my RHEL test servers and one GIG without swap is simply not enough for Katello. Mongo and Tomcat were unlucky in the russian roulette. It works even when I have the update repo disabled. So my second point is not valid - updated wiki. https://fedorahosted.org/katello/wiki/Install?action=diff&version=101 On the other hand - updates are highly recommended for users. I left the warning there. LZ On Thu, Dec 08, 2011 at 07:42:56AM -0500, Bryan Kearney wrote: > On 12/08/2011 04:22 AM, Lukas Zapletal wrote: > >Hello, > > > >today I discovered few facts. > > > >If you try to install Katello on Fedora 15 with updates disabled, you > >get an dependency error (Puppet 2.6 required, but Fedora GOLD contains > >older version). > > > >On RHEL 6.1 with updates turned off there are two different issues. > >Although you are able to install, I am running into two issues: > > > >- if I restart pulp-sever mongodb segfaults which causes that lock file > > is blocking pulp from starting > > > >- after some decent amount of requests (our system test cli suite) > > tomcat falls down without any warning or error - the process just got > > killed > > > >With updates applied I don't have such issues. > > > >Instead of crying here, I am editing our wiki and putting a note for our > >users to update their systems before installing. > > I agree.. I think this is fine. > > > > >https://fedorahosted.org/katello/wiki/Install > > I updated > > https://fedorahosted.org/candlepin/wiki/headpin/Install > > as well > > > > >I don't know why, but I am humming "Sign of Punk" song this morning ;-) > > > >http://www.youtube.com/watch?v=ZuYx_z9KU68 > > > > Too early. > > -- bk > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From inecas at redhat.com Thu Dec 8 13:15:27 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Thu, 08 Dec 2011 14:15:27 +0100 Subject: [katello-devel] Import manifest orchestration reworked In-Reply-To: <20111208114421.GC8512@lzapx.brq.redhat.com> References: <4EDFA323.10701@redhat.com> <20111208114421.GC8512@lzapx.brq.redhat.com> Message-ID: <4EE0B86F.9030804@redhat.com> On 12/08/2011 12:44 PM, Lukas Zapletal wrote: > So you also solved the issue when uebercert was blocking whole thing? No > need of uebercert deletion anymore? It seems no deed for that. Deleting works just fine now (at least is seems). > > > From my point of view, it looks very good. > > This change needs intensive testing it can have consequences we cannot > currently think of. Bryan does it worth to include it in the drop? > Because it's a good improvement. Fully agree on that. It's quite a big change (although it changes the interface only slightly). On the other hand it solves problems we already know we had before (importing manifests for more organizations, especially when one organization had already some stuff promoted, importing manifest after creating uebercert etc.). I've asked Garik to run QA tests again this branch and it looks ok there as well. From my point of view I would like to put it into the drop, partly because then it will get even more testing. But depends what consequences would have introducing some new bug into this drop. -- Ivan > > Good work! > > LZ > > On Wed, Dec 07, 2011 at 06:32:19PM +0100, Ivan Necas wrote: >> Hi all, >> >> Few last days I was working on the rework of orchestration of >> manifest import and products promotion. Reasons why it needed to be >> changed and what the solution looks like are described on a wiki >> page [1]. In short, with this changes we don't need to change >> Candlepin content when promoting products/repos. >> >> Since this is quite a big change in our orchestration model, I would >> like you to review the changes and provide a feedback/comments to >> the current implementation. The changes are available in a remote >> branch *reimport-manifest*. >> >> [1] - https://fedorahosted.org/katello/wiki/CandlepinEnvironmentsIntegration >> >> PS: this changes also make our system tests green again:) >> >> -- Ivan >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel > -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From jweiss at redhat.com Thu Dec 8 13:35:02 2011 From: jweiss at redhat.com (Jeff Weiss) Date: Thu, 08 Dec 2011 08:35:02 -0500 Subject: [katello-devel] Katello SSL issues In-Reply-To: <20111207093540.GB19321@lzapx.brq.redhat.com> References: <20111206161417.GC3471@lzapx.brq.redhat.com> <1323206997.10548.13.camel@blinky.rdu.redhat.com> <20111207093540.GB19321@lzapx.brq.redhat.com> Message-ID: <1323351302.25307.3.camel@jweiss> On Wed, 2011-12-07 at 10:35 +0100, Lukas Zapletal wrote: > Jeff, those are good points. > > It looks like a sprint task for Pulp team to get it working. Now, > correct me if I am wrong, but our main idea is to have one "main" > (Katello) CA that will sign all the others. > > LZ > Lukas, I'm not sure what the relationship between the Katello and pulp/cp certs should be. Since katello puts data in and takes data out of pulp/cp, it seems reasonable to me that Katello's CA would sign the others. Just want to make sure the use case is known where one more CA will be in the chain (the customer's CA). -jeff > On Tue, Dec 06, 2011 at 04:29:57PM -0500, Jeff Weiss wrote: > > On Tue, 2011-12-06 at 17:14 +0100, Lukas Zapletal wrote: > > > Hey, > > > > > > as I wrote yesterday, I accepted Tomas' patch. We have now the tool and > > > few Puppet steps which generate CA cert and Apache http/Qpid certs, but > > > they are unused right now. > > > > > > I would recommend to postpone work on SSL certs until end of next week. > > > Then we can re-enable them and start with tuning and testing. I expect > > > some changes on the Candlepin/Pulp side, because the plan is to have > > > "the big" Katello CA and have all certificates signed with it. That > > > could possibly cause troubles (instead checking one CA clients will need > > > to check the whole chain). Instead of big bang, I would stick with > > > step-by-step approach. We can start with Pulp as I think the integration > > > should be easier and then move on with Candlepin (that's gonna hurt). > > > > > > By the way there is/was an issue with our SSL generation due to race > > > condition/working directory issue in Puppet. I hope this is solved now, > > > but until packages get built, please make sure you start > > > katello-configure in the /root directory: > > > > > > pushd /root && katello-configure && popd > > > > > > https://bugzilla.redhat.com/show_bug.cgi?id=760280 > > > > > > Please not there is also another bug, if you do not have FQDN properly > > > set, Puppet "guess" it from /etc/resolv.conf which cause troubles. Fix > > > have been commited moments ago: > > > > > > https://bugzilla.redhat.com/show_bug.cgi?id=760265 > > > > > > In short - katello-configure-0.1.26-1 should be fine. Please report > > > here. > > > > > > > I've done a lot of testing with SSL already - browsers don't like > > self-signed certs, and automated testing with Selenium can't get past > > the warning by itself. So I started using my own CA to sign the > > candlepin certs. > > > > With a bit of scripting to sign and copy the certs to the right places, > > the Katello UI works fine. With a bit more configuring of apache, RHSM > > clients can register. > > > > But I got totally stuck where Pulp cannot validate the client certs > > coming from RHSM. I believe the problem may be that pulp will only > > check against one cert, not the whole chain (it fails if there's more > > than 1 in the chain). The katello installer sets up pulp to point at > > Candlepin's cert but doesn't include my CA in the chain. But it's more > > than that, I suspect. I don't think pulp is even capable of pointing at > > multiple CA files, and it doesn't seem to accept them concatenated > > together the way most SSL libs do. > > > > Jay Dobies was helping me out with this a few weeks ago but we did not > > finish getting to the root of the problem. > > > > > > -- > > Jeff Weiss > > Principal Quality Assurance Engineer > > jweiss at redhat.com > > (919)754-4178 > > > > _______________________________________________ > > katello-devel mailing list > > katello-devel at redhat.com > > https://www.redhat.com/mailman/listinfo/katello-devel > From lzap at redhat.com Thu Dec 8 13:36:58 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 8 Dec 2011 14:36:58 +0100 Subject: [katello-devel] Import manifest orchestration reworked In-Reply-To: <4EE0B86F.9030804@redhat.com> References: <4EDFA323.10701@redhat.com> <20111208114421.GC8512@lzapx.brq.redhat.com> <4EE0B86F.9030804@redhat.com> Message-ID: <20111208133657.GF8512@lzapx.brq.redhat.com> > From my point of view I would like to put it into the drop, partly > because then it will get even more testing. But depends what > consequences would have introducing some new bug into this drop. > I fully agree with this opinion. Less dupe-bugzillas on this topic, more testing and still - if we find it broke something important, we can revert those four patches in the branch. But as Garik already did preliminary testing and it felt good we should be fine. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Thu Dec 8 14:55:50 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 8 Dec 2011 15:55:50 +0100 Subject: [katello-devel] katello-configure-0.1.27-1 broken Message-ID: <20111208145550.GA13220@lzapx.brq.redhat.com> Hey, I just broke katello-configure. Version katello-configure-0.1.27-1 fails before puppet starts. The workaround is to edit: /usr/share/katello/puppet/modules/katello/manifests/params.pp and delete one of the definitions of $migrate_log variables (does not matter which one). It's around line 28. Fix was pushed, gibson is working on a new build. "This cannot hurt," I thought. And I was wrong. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From bkearney at redhat.com Thu Dec 8 14:57:39 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Thu, 08 Dec 2011 09:57:39 -0500 (EST) Subject: [katello-devel] Entitlement team Sprint 36 Review Message-ID: The following is a new meeting request: Subject: Entitlement team Sprint 36 Review Organizer: "Bryan Kearney" Location: "UNC - RDU" Resources: "UNC - RDU" (Site: RDU, Building: Varsity, Floor: 4, Room: 421A) Time: Wednesday, January 11, 2012, 9:00:00 AM - 10:00:00 AM GMT -05:00 US/Canada Eastern Invitees: chainsaw at redhat.com; entitlement-team-list at redhat.com; katello-devel at redhat.com *~*~*~*~*~*~*~*~*~* -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: meeting.ics Type: text/calendar Size: 2931 bytes Desc: not available URL: From bbuckingham at redhat.com Thu Dec 8 15:13:49 2011 From: bbuckingham at redhat.com (Brad Buckingham) Date: Thu, 08 Dec 2011 10:13:49 -0500 Subject: [katello-devel] system package actions - branch merged to master Message-ID: <4EE0D42D.4010104@redhat.com> Team, I've just merged the sys-packages branch to master. With this merge, a user may perform the following from the UI: - Go to Systems -> [pick a system] -> Packages 1. Schedule ('immediate') Package install, remove and update actions for the system 2. Schedule ('immediate') Package Group install/remove actions for the system 3. Schedule ('immediate') 'update all' for the system (e.g. similar to 'yum update') Once the action is scheduled, the UI will poll for status of those actions providing an indicator to the user that the action is in progress, updating it if it completes or fails. Things to note: - In order for the above actions to be executed properly. The client system must have the katello-agent installed. This is a new RPM that will be made available soon in repos soon; however, developers can build from within git, if needed. Stay tuned for email when the agent is available in the repos. Details on the agent can also be found at: https://fedorahosted.org/katello/wiki/KatelloAgent . - When scheduling an action using the 'Enter Selected Content' box on the Packages pane, the user may enter multiple package or package group names, separated by a comma (e.g. 'xterm, zsh, zip'). Impact to developers: - This feature includes modifications to the task model for storing the status of these actions. As a result, you will need to at least run a 'rake db:migrate'. Issues: - Currently, when the Packages pane loads, the UI does not display an actions that are currently in progress. We are working on enhancements to address this and hope to have that in very soon; however, making this code available now since base functionality is available. If you have any questions, please do let us know. thanks, Brad From inecas at redhat.com Thu Dec 8 15:21:25 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Thu, 08 Dec 2011 16:21:25 +0100 Subject: [katello-devel] system package actions - branch merged to master In-Reply-To: <4EE0D42D.4010104@redhat.com> References: <4EE0D42D.4010104@redhat.com> Message-ID: <4EE0D5F5.3090000@redhat.com> Nice, What pulp version do we need to run this? I am on pulp-0.0.244-5.fc15.noarch and I'm getting: Pulp::Consumer: 404 Resource Not Found not found (POST /pulp/api/consumers /e60b0f05-5069-433c-a8d6-6b4f6a9f88ab/updatepackages/): not found when I try to install new package. -- Ivan On 12/08/2011 04:13 PM, Brad Buckingham wrote: > Team, > > I've just merged the sys-packages branch to master. > > With this merge, a user may perform the following from the UI: > > - Go to Systems -> [pick a system] -> Packages > > 1. Schedule ('immediate') Package install, remove and update > actions for the system > 2. Schedule ('immediate') Package Group install/remove actions for > the system > 3. Schedule ('immediate') 'update all' for the system (e.g. > similar to 'yum update') > > Once the action is scheduled, the UI will poll for status of those > actions providing an indicator to the user that the action is in > progress, updating it if it completes or fails. > > Things to note: > > - In order for the above actions to be executed properly. The client > system must have the katello-agent installed. This is a new RPM that > will be made available soon in repos soon; however, developers can > build from within git, if needed. Stay tuned for email when the agent > is available in the repos. Details on the agent can also be found at: > https://fedorahosted.org/katello/wiki/KatelloAgent . > > - When scheduling an action using the 'Enter Selected Content' box on > the Packages pane, the user may enter multiple package or package > group names, separated by a comma (e.g. 'xterm, zsh, zip'). > > Impact to developers: > > - This feature includes modifications to the task model for storing > the status of these actions. As a result, you will need to at least > run a 'rake db:migrate'. > > Issues: > > - Currently, when the Packages pane loads, the UI does not display an > actions that are currently in progress. We are working on > enhancements to address this and hope to have that in very soon; > however, making this code available now since base functionality is > available. > > If you have any questions, please do let us know. > > thanks, > Brad > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From bbuckingham at redhat.com Thu Dec 8 15:23:39 2011 From: bbuckingham at redhat.com (Brad Buckingham) Date: Thu, 08 Dec 2011 10:23:39 -0500 Subject: [katello-devel] system package actions - branch merged to master In-Reply-To: <4EE0D5F5.3090000@redhat.com> References: <4EE0D42D.4010104@redhat.com> <4EE0D5F5.3090000@redhat.com> Message-ID: <4EE0D67B.6010600@redhat.com> The 'updatepackages' is a fairly recent addition to the pulp API. If running a version older than the following, you may see the error below: pulp-0.0.254-1.fc15.noarch thanks, Brad On 12/08/2011 10:21 AM, Ivan Ne?as wrote: > Nice, > > What pulp version do we need to run this? I am on > pulp-0.0.244-5.fc15.noarch and I'm getting: > > Pulp::Consumer: 404 Resource Not Found not found (POST > /pulp/api/consumers > /e60b0f05-5069-433c-a8d6-6b4f6a9f88ab/updatepackages/): not found > > when I try to install new package. > > -- Ivan > On 12/08/2011 04:13 PM, Brad Buckingham wrote: >> Team, >> >> I've just merged the sys-packages branch to master. >> >> With this merge, a user may perform the following from the UI: >> >> - Go to Systems -> [pick a system] -> Packages >> >> 1. Schedule ('immediate') Package install, remove and update actions >> for the system >> 2. Schedule ('immediate') Package Group install/remove actions for >> the system >> 3. Schedule ('immediate') 'update all' for the system (e.g. similar >> to 'yum update') >> >> Once the action is scheduled, the UI will poll for status of those >> actions providing an indicator to the user that the action is in >> progress, updating it if it completes or fails. >> >> Things to note: >> >> - In order for the above actions to be executed properly. The client >> system must have the katello-agent installed. This is a new RPM that >> will be made available soon in repos soon; however, developers can >> build from within git, if needed. Stay tuned for email when the agent >> is available in the repos. Details on the agent can also be found at: >> https://fedorahosted.org/katello/wiki/KatelloAgent . >> >> - When scheduling an action using the 'Enter Selected Content' box on >> the Packages pane, the user may enter multiple package or package >> group names, separated by a comma (e.g. 'xterm, zsh, zip'). >> >> Impact to developers: >> >> - This feature includes modifications to the task model for storing >> the status of these actions. As a result, you will need to at least >> run a 'rake db:migrate'. >> >> Issues: >> >> - Currently, when the Packages pane loads, the UI does not display an >> actions that are currently in progress. We are working on >> enhancements to address this and hope to have that in very soon; >> however, making this code available now since base functionality is >> available. >> >> If you have any questions, please do let us know. >> >> thanks, >> Brad >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel > > From lzap at redhat.com Thu Dec 8 15:26:04 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 8 Dec 2011 16:26:04 +0100 Subject: [katello-devel] katello-configure-0.1.27-1 broken In-Reply-To: <20111208145550.GA13220@lzapx.brq.redhat.com> References: <20111208145550.GA13220@lzapx.brq.redhat.com> Message-ID: <20111208152603.GB13220@lzapx.brq.redhat.com> Fixed version is: katello-configure-0.1.29-1 For older version I am correcting the the file path: /usr/share/katello/install/puppet/modules/katello/manifests/params.pp delete the double definition on line 28 (migrate_log) AND /usr/share/katello/install/puppet/modules/katello/manifests/config.pp Change line command => "/usr/bin/env > ${katello::params::dbenv_log}", to command => "/usr/bin/env > ${katello::params::db_env_log}", Uh. On Thu, Dec 08, 2011 at 03:55:50PM +0100, Lukas Zapletal wrote: > Hey, > > I just broke katello-configure. Version katello-configure-0.1.27-1 fails > before puppet starts. > > The workaround is to edit: > > /usr/share/katello/puppet/modules/katello/manifests/params.pp > > and delete one of the definitions of $migrate_log variables (does not > matter which one). It's around line 28. > > Fix was pushed, gibson is working on a new build. > > "This cannot hurt," I thought. And I was wrong. > > -- > Later, > > Lukas Zapletal | E32E400A > RHN Satellite Engineering > Red Hat Czech s.r.o. Brno > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From cperry at redhat.com Thu Dec 8 15:27:15 2011 From: cperry at redhat.com (Cliff Perry) Date: Thu, 08 Dec 2011 10:27:15 -0500 Subject: [katello-devel] Fwd: [katello-commits] [katello] puppet - force creation of the candlepin symlink Message-ID: <4EE0D753.1060107@redhat.com> FYI Lukas - I think this commit fixes: Repeat runs of katello-configure fail because of existing file https://bugzilla.redhat.com/show_bug.cgi?id=758509 Cliff -------- Original Message -------- Subject: [katello-commits] [katello] puppet - force creation of the candlepin symlink Date: Thu, 8 Dec 2011 15:20:09 +0000 (UTC) From: lzap Reply-To: katello-commits at lists.fedorahosted.org To: katello-commits at lists.fedorahosted.org commit 36756bef0f99b361c6ec682d027607350f83bd1a Author: Lukas Zapletal Date: Thu Dec 8 16:19:40 2011 +0100 puppet - force creation of the candlepin symlink puppet/modules/pulp/manifests/config.pp | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) --- diff --git a/puppet/modules/pulp/manifests/config.pp b/puppet/modules/pulp/manifests/config.pp index 99bedba..2752626 100644 --- a/puppet/modules/pulp/manifests/config.pp +++ b/puppet/modules/pulp/manifests/config.pp @@ -23,7 +23,7 @@ class pulp::config { } exec { "set candlepin crl file": - command => "/usr/bin/openssl x509 -in '$pulp::params::ssl_certificate_file' -hash -noout | /usr/bin/xargs -I{} /bin/ln -s '$candlepin::params::crl_file' '$pulp::params::crl_location/{}.r0'", + command => "/usr/bin/openssl x509 -in '$pulp::params::ssl_certificate_file' -hash -noout | /usr/bin/xargs -I{} /bin/ln -sf '$candlepin::params::crl_file' '$pulp::params::crl_location/{}.r0'", require => Class["candlepin::config"], } _______________________________________________ katello-commits mailing list katello-commits at lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/katello-commits From inecas at redhat.com Thu Dec 8 16:19:11 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Thu, 08 Dec 2011 17:19:11 +0100 Subject: [katello-devel] system package actions - branch merged to master In-Reply-To: <4EE0D67B.6010600@redhat.com> References: <4EE0D42D.4010104@redhat.com> <4EE0D5F5.3090000@redhat.com> <4EE0D67B.6010600@redhat.com> Message-ID: <4EE0E37F.5080202@redhat.com> Another note: Make sure you have the very same version of gofer on the client and the server, otherwise you get: 2011-12-08 17:02:08,989 [WARNING][9aa0a9dc-f2a2-471b-b69a-07694aaef821] valid() @ consumer.py:234 - {9aa0a9dc-f2a2-471b-b69a-07694aaef821} version mismatch (discarded): when you try install a package remotely -- Ivan On 12/08/2011 04:23 PM, Brad Buckingham wrote: > The 'updatepackages' is a fairly recent addition to the pulp API. If > running a version older than the following, you may see the error below: > > pulp-0.0.254-1.fc15.noarch > > thanks, > Brad > > On 12/08/2011 10:21 AM, Ivan Ne?as wrote: >> Nice, >> >> What pulp version do we need to run this? I am on >> pulp-0.0.244-5.fc15.noarch and I'm getting: >> >> Pulp::Consumer: 404 Resource Not Found not found (POST >> /pulp/api/consumers >> /e60b0f05-5069-433c-a8d6-6b4f6a9f88ab/updatepackages/): not found >> >> when I try to install new package. >> >> -- Ivan >> On 12/08/2011 04:13 PM, Brad Buckingham wrote: >>> Team, >>> >>> I've just merged the sys-packages branch to master. >>> >>> With this merge, a user may perform the following from the UI: >>> >>> - Go to Systems -> [pick a system] -> Packages >>> >>> 1. Schedule ('immediate') Package install, remove and update actions >>> for the system >>> 2. Schedule ('immediate') Package Group install/remove actions for >>> the system >>> 3. Schedule ('immediate') 'update all' for the system (e.g. similar >>> to 'yum update') >>> >>> Once the action is scheduled, the UI will poll for status of those >>> actions providing an indicator to the user that the action is in >>> progress, updating it if it completes or fails. >>> >>> Things to note: >>> >>> - In order for the above actions to be executed properly. The client >>> system must have the katello-agent installed. This is a new RPM that >>> will be made available soon in repos soon; however, developers can >>> build from within git, if needed. Stay tuned for email when the >>> agent is available in the repos. Details on the agent can also be >>> found at: https://fedorahosted.org/katello/wiki/KatelloAgent . >>> >>> - When scheduling an action using the 'Enter Selected Content' box >>> on the Packages pane, the user may enter multiple package or package >>> group names, separated by a comma (e.g. 'xterm, zsh, zip'). >>> >>> Impact to developers: >>> >>> - This feature includes modifications to the task model for storing >>> the status of these actions. As a result, you will need to at least >>> run a 'rake db:migrate'. >>> >>> Issues: >>> >>> - Currently, when the Packages pane loads, the UI does not display >>> an actions that are currently in progress. We are working on >>> enhancements to address this and hope to have that in very soon; >>> however, making this code available now since base functionality is >>> available. >>> >>> If you have any questions, please do let us know. >>> >>> thanks, >>> Brad >>> >>> _______________________________________________ >>> katello-devel mailing list >>> katello-devel at redhat.com >>> https://www.redhat.com/mailman/listinfo/katello-devel >> >> > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From bkearney at redhat.com Thu Dec 8 16:30:07 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Thu, 08 Dec 2011 11:30:07 -0500 Subject: [katello-devel] system package actions - branch merged to master In-Reply-To: <4EE0E37F.5080202@redhat.com> References: <4EE0D42D.4010104@redhat.com> <4EE0D5F5.3090000@redhat.com> <4EE0D67B.6010600@redhat.com> <4EE0E37F.5080202@redhat.com> Message-ID: <4EE0E60F.1090907@redhat.com> On 12/08/2011 11:19 AM, Ivan Ne?as wrote: > Another note: > > Make sure you have the very same version of gofer on the client and the > server, otherwise you get: > > 2011-12-08 17:02:08,989 [WARNING][9aa0a9dc-f2a2-471b-b69a-07694aaef821] > valid() @ consumer.py:234 - {9aa0a9dc-f2a2-471b-b69a-07694aaef821} > version mismatch (discarded): > > when you try install a package remotely > > -- Ivan > I assume there is a bug open for this? This will not be acceptable long term. -- bk From mmccune at redhat.com Thu Dec 8 16:53:22 2011 From: mmccune at redhat.com (Mike McCune) Date: Thu, 08 Dec 2011 08:53:22 -0800 Subject: [katello-devel] katello-agent src dir moved Message-ID: <4EE0EB82.8060009@redhat.com> just a heads up, I moved the src dir from: katello/client to: katello/agent Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From jsherril at redhat.com Thu Dec 8 17:01:10 2011 From: jsherril at redhat.com (Justin Sherrill) Date: Thu, 08 Dec 2011 12:01:10 -0500 Subject: [katello-devel] Non-ar search discussion Message-ID: <4EE0ED56.6040203@redhat.com> Hi All, I've been tasked with looking into a mechanism to do non-ar search. While amos had leaned towards a search proxy, I'm more inclined to utilize a full text search server. I've put some of the pros/cons here: https://fedorahosted.org/katello/wiki/KatelloSearchRevamp Along with a list of things that need to be searched. Thoughts? Comments? -Justin From inecas at redhat.com Thu Dec 8 17:08:29 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Thu, 08 Dec 2011 18:08:29 +0100 Subject: [katello-devel] Import manifest orchestration reworked In-Reply-To: <20111208133657.GF8512@lzapx.brq.redhat.com> References: <4EDFA323.10701@redhat.com> <20111208114421.GC8512@lzapx.brq.redhat.com> <4EE0B86F.9030804@redhat.com> <20111208133657.GF8512@lzapx.brq.redhat.com> Message-ID: <4EE0EF0D.8070805@redhat.com> On 12/08/2011 02:36 PM, Lukas Zapletal wrote: >> From my point of view I would like to put it into the drop, partly >> because then it will get even more testing. But depends what >> consequences would have introducing some new bug into this drop. >> > I fully agree with this opinion. Less dupe-bugzillas on this topic, more > testing and still - if we find it broke something important, we can > revert those four patches in the branch. But as Garik already did > preliminary testing and it felt good we should be fine. > Merged into master now. -- Ivan From lzap at redhat.com Thu Dec 8 17:11:46 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 8 Dec 2011 18:11:46 +0100 Subject: [katello-devel] Import manifest orchestration reworked In-Reply-To: <4EE0EF0D.8070805@redhat.com> References: <4EDFA323.10701@redhat.com> <20111208114421.GC8512@lzapx.brq.redhat.com> <4EE0B86F.9030804@redhat.com> <20111208133657.GF8512@lzapx.brq.redhat.com> <4EE0EF0D.8070805@redhat.com> Message-ID: <20111208171146.GC13220@lzapx.brq.redhat.com> > Merged into master now. > Cool. Do NOT leave the office today ;-) -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From jrist at redhat.com Thu Dec 8 17:21:27 2011 From: jrist at redhat.com (Jason Rist) Date: Thu, 08 Dec 2011 10:21:27 -0700 Subject: [katello-devel] Non-ar search discussion In-Reply-To: <4EE0ED56.6040203@redhat.com> References: <4EE0ED56.6040203@redhat.com> Message-ID: <4EE0F217.6030409@redhat.com> On Thu 08 Dec 2011 10:01:10 AM MST, Justin Sherrill wrote: > Hi All, > > I've been tasked with looking into a mechanism to do non-ar search. > While amos had leaned towards a search proxy, I'm more inclined to > utilize a full text search server. I've put some of the pros/cons here: > > https://fedorahosted.org/katello/wiki/KatelloSearchRevamp > > Along with a list of things that need to be searched. > > Thoughts? Comments? > > -Justin > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel Huge +1 on full text server. This is territory where lots of things have been done and should be relatively straight forward. -J -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From mmccune at redhat.com Thu Dec 8 17:39:16 2011 From: mmccune at redhat.com (Mike McCune) Date: Thu, 08 Dec 2011 09:39:16 -0800 Subject: [katello-devel] Non-ar search discussion In-Reply-To: <4EE0F217.6030409@redhat.com> References: <4EE0ED56.6040203@redhat.com> <4EE0F217.6030409@redhat.com> Message-ID: <4EE0F644.6080903@redhat.com> On 12/08/2011 09:21 AM, Jason Rist wrote: > On Thu 08 Dec 2011 10:01:10 AM MST, Justin Sherrill wrote: >> Hi All, >> >> I've been tasked with looking into a mechanism to do non-ar search. >> While amos had leaned towards a search proxy, I'm more inclined to >> utilize a full text search server. I've put some of the pros/cons here: >> >> https://fedorahosted.org/katello/wiki/KatelloSearchRevamp >> >> Along with a list of things that need to be searched. >> >> Thoughts? Comments? >> >> -Justin >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel > > Huge +1 on full text server. This is territory where lots of things > have been done and should be relatively straight forward. > > -J I think we could end up re-using Spacewalk's search server if so (not my idea, I think JesusR came up with that one). Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From ehelms at redhat.com Thu Dec 8 17:42:30 2011 From: ehelms at redhat.com (Eric Helms) Date: Thu, 08 Dec 2011 12:42:30 -0500 Subject: [katello-devel] Non-ar search discussion In-Reply-To: <4EE0F217.6030409@redhat.com> References: <4EE0ED56.6040203@redhat.com> <4EE0F217.6030409@redhat.com> Message-ID: <4EE0F706.8030807@redhat.com> On 12/08/2011 12:21 PM, Jason Rist wrote: > On Thu 08 Dec 2011 10:01:10 AM MST, Justin Sherrill wrote: >> Hi All, >> >> I've been tasked with looking into a mechanism to do non-ar search. >> While amos had leaned towards a search proxy, I'm more inclined to >> utilize a full text search server. I've put some of the pros/cons here: >> >> https://fedorahosted.org/katello/wiki/KatelloSearchRevamp >> >> Along with a list of things that need to be searched. >> >> Thoughts? Comments? >> >> -Justin >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel > Huge +1 on full text server. This is territory where lots of things > have been done and should be relatively straight forward. > > -J > +1 to robust solutions where people have already solved this problem and combine both backend service searching and Katello model searching into a single interface. Are there any performance / resource tradeoffs between the two methods? (e.g. setup cost, more/less memory usage, dependencies) Would the indexing delay only affect the availability of searching up to date data or have an overall performance impact on the app? - Eric H. From jsherril at redhat.com Thu Dec 8 17:50:39 2011 From: jsherril at redhat.com (Justin Sherrill) Date: Thu, 08 Dec 2011 12:50:39 -0500 Subject: [katello-devel] Non-ar search discussion In-Reply-To: <4EE0F706.8030807@redhat.com> References: <4EE0ED56.6040203@redhat.com> <4EE0F217.6030409@redhat.com> <4EE0F706.8030807@redhat.com> Message-ID: <4EE0F8EF.3030504@redhat.com> On 12/08/2011 12:42 PM, Eric Helms wrote: > On 12/08/2011 12:21 PM, Jason Rist wrote: >> On Thu 08 Dec 2011 10:01:10 AM MST, Justin Sherrill wrote: >>> Hi All, >>> >>> I've been tasked with looking into a mechanism to do non-ar search. >>> While amos had leaned towards a search proxy, I'm more inclined to >>> utilize a full text search server. I've put some of the pros/cons >>> here: >>> >>> https://fedorahosted.org/katello/wiki/KatelloSearchRevamp >>> >>> Along with a list of things that need to be searched. >>> >>> Thoughts? Comments? >>> >>> -Justin >>> >>> _______________________________________________ >>> katello-devel mailing list >>> katello-devel at redhat.com >>> https://www.redhat.com/mailman/listinfo/katello-devel >> Huge +1 on full text server. This is territory where lots of things >> have been done and should be relatively straight forward. >> >> -J >> > +1 to robust solutions where people have already solved this problem > and combine both backend service searching and Katello model searching > into a single interface. > > Are there any performance / resource tradeoffs between the two > methods? (e.g. setup cost, more/less memory usage, dependencies) > > Would the indexing delay only affect the availability of searching up > to date data or have an overall performance impact on the app? > I'm not sure that the setup cost would be any more for either method. The fulltext search would most likely use more memory, have more dependencies (depending on a number of factors). and require more disk space. None of these things are deal breakers though, especially given the advantages. The indexing delay should not have much effect on performance just a 'lag' on being able to search for a new or updated item. Also, as i said in the wiki, I think the only noticeable delay would be when a repo is synced and we try to re-index the packages and errata. Generally any update or create would be handled in the background as the create or update occurs. So generally before the user would be able to search for the updated item, it would be updated. > - Eric H. > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel From jsherril at redhat.com Thu Dec 8 18:08:45 2011 From: jsherril at redhat.com (Justin Sherrill) Date: Thu, 08 Dec 2011 13:08:45 -0500 Subject: [katello-devel] Non-ar search discussion In-Reply-To: <4EE0F644.6080903@redhat.com> References: <4EE0ED56.6040203@redhat.com> <4EE0F217.6030409@redhat.com> <4EE0F644.6080903@redhat.com> Message-ID: <4EE0FD2D.1020203@redhat.com> On 12/08/2011 12:39 PM, Mike McCune wrote: > On 12/08/2011 09:21 AM, Jason Rist wrote: >> On Thu 08 Dec 2011 10:01:10 AM MST, Justin Sherrill wrote: >>> Hi All, >>> >>> I've been tasked with looking into a mechanism to do non-ar search. >>> While amos had leaned towards a search proxy, I'm more inclined to >>> utilize a full text search server. I've put some of the pros/cons >>> here: >>> >>> https://fedorahosted.org/katello/wiki/KatelloSearchRevamp >>> >>> Along with a list of things that need to be searched. >>> >>> Thoughts? Comments? >>> >>> -Justin >>> >>> _______________________________________________ >>> katello-devel mailing list >>> katello-devel at redhat.com >>> https://www.redhat.com/mailman/listinfo/katello-devel >> >> Huge +1 on full text server. This is territory where lots of things >> have been done and should be relatively straight forward. >> >> -J > > I think we could end up re-using Spacewalk's search server if so (not > my idea, I think JesusR came up with that one). > > Mike We could, although since the spacewalk search server was created there are a number of alternatives that present a simple rest interface for indexing and storing data. I believe it was built on the idea that it had access to the spacewalk database which I'm not sure we would want for our backend systems. John pointed me to this one: http://www.elasticsearch.org/ Many rails people seem to use this one: http://wiki.apache.org/solr/ or http://sphinxsearch.com/ These are more 'prebuilt' servers that have defined ways of getting data into them. If we were to go this route, we would most likely use background tasks within katello to retrieve data from the backend systems and push them to the server when needed. Most of these have rails plugins that will automatically handle the updating of models in the search server for AR data. Elastic search seems the simplest to get started with simply because its just pushing json structures to the server, and constructing queries from json structures. -Justin From jortel at redhat.com Thu Dec 8 19:40:32 2011 From: jortel at redhat.com (Jeff Ortel) Date: Thu, 08 Dec 2011 13:40:32 -0600 Subject: [katello-devel] system package actions - branch merged to master In-Reply-To: <4EE0E60F.1090907@redhat.com> References: <4EE0D42D.4010104@redhat.com> <4EE0D5F5.3090000@redhat.com> <4EE0D67B.6010600@redhat.com> <4EE0E37F.5080202@redhat.com> <4EE0E60F.1090907@redhat.com> Message-ID: <4EE112B0.10409@redhat.com> On 12/08/2011 10:30 AM, Bryan Kearney wrote: > On 12/08/2011 11:19 AM, Ivan Ne??as wrote: >> Another note: >> >> Make sure you have the very same version of gofer on the client and the >> server, otherwise you get: >> >> 2011-12-08 17:02:08,989 [WARNING][9aa0a9dc-f2a2-471b-b69a-07694aaef821] >> valid() @ consumer.py:234 - {9aa0a9dc-f2a2-471b-b69a-07694aaef821} >> version mismatch (discarded): >> >> when you try install a package remotely >> >> -- Ivan >> > > I assume there is a bug open for this? This will not be acceptable long term. This indicates a mismatch in gofer's communication protocol version which has been (and should continue to) be very static. As new features are added, new properties are sometimes added to the protocol. When this happens, new properties are defaulted and don't require a protocol version change. Basically this protocol is normally backwards compatible. In a recent gofer release, I chose to bump the version to reflect protocol changes over the past year. Really just a pre-v1 administrative thing. Sorry for the inconvenience. Historically, we haven't needed to have the *very* same version of gofer running on both the client and server and I don't see that changing going forward (Long Term). > > -- bk > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel From mmccune at redhat.com Fri Dec 9 01:13:57 2011 From: mmccune at redhat.com (Mike McCune) Date: Thu, 08 Dec 2011 17:13:57 -0800 Subject: [katello-devel] errors with latest packages Message-ID: <4EE160D5.6070702@redhat.com> # katello-configure Starting Katello configuration The top-level log file is [/var/log/katello/katello-configure-20111209-091609/main.log] err: /Stage[main]/Certs::Config/Exec[create-nss-db]/returns: change from notrun to 0 failed: /bin/rm -f /etc/pki/katello/nssdb//*; certutil -N -d '/etc/pki/katello/nssdb/' -f '/etc/katello/nss_db_password-file'; certutil -A -d '/etc/pki/katello/nssdb/' -n 'ca' -t 'TCu,Cu,Tuw' -a -i '/usr/share/katello/KATELLO-TRUSTED-SSL-CERT'; certutil -A -d '/etc/pki/katello/nssdb/' -n 'broker' -t ',,' -a -i '/etc/pki/tls/certs/qpid-broker.crt'; certutil -A -d '/etc/pki/katello/nssdb/' -n 'tomcat' -t ',,' -a -i '/etc/pki/tls/certs/httpd-ssl.crt' returned 255 instead of one of [0] at /usr/share/katello/install/puppet/modules/certs/manifests/config.pp:184 katello-0.1.135-2.el6.x86_64 katello-all-0.1.135-2.el6.x86_64 katello-certs-tools-1.0.1-2.el6.noarch katello-cli-0.1.23-3.el6.noarch katello-cli-common-0.1.23-3.el6.noarch katello-common-0.1.135-2.el6.noarch katello-configure-0.1.31-2.el6.noarch katello-glue-candlepin-0.1.135-2.el6.x86_64 katello-glue-foreman-0.1.135-2.el6.x86_64 katello-glue-pulp-0.1.135-2.el6.x86_64 katello-qpid-broker-key-pair-1.0-1.noarch katello-trusted-ssl-cert-1.0-1.noarch Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From mmccune at redhat.com Fri Dec 9 04:50:21 2011 From: mmccune at redhat.com (Mike McCune) Date: Thu, 08 Dec 2011 20:50:21 -0800 Subject: [katello-devel] errors with latest packages In-Reply-To: <4EE160D5.6070702@redhat.com> References: <4EE160D5.6070702@redhat.com> Message-ID: <4EE1938D.8050808@redhat.com> On 12/08/2011 05:13 PM, Mike McCune wrote: > # katello-configure > Starting Katello configuration > The top-level log file is > [/var/log/katello/katello-configure-20111209-091609/main.log] > err: /Stage[main]/Certs::Config/Exec[create-nss-db]/returns: change from > notrun to 0 failed: /bin/rm -f /etc/pki/katello/nssdb//*; certutil -N -d > '/etc/pki/katello/nssdb/' -f '/etc/katello/nss_db_password-file'; > certutil -A -d '/etc/pki/katello/nssdb/' -n 'ca' -t 'TCu,Cu,Tuw' -a -i > '/usr/share/katello/KATELLO-TRUSTED-SSL-CERT'; certutil -A -d > '/etc/pki/katello/nssdb/' -n 'broker' -t ',,' -a -i > '/etc/pki/tls/certs/qpid-broker.crt'; certutil -A -d > '/etc/pki/katello/nssdb/' -n 'tomcat' -t ',,' -a -i > '/etc/pki/tls/certs/httpd-ssl.crt' returned 255 instead of one of [0] at > /usr/share/katello/install/puppet/modules/certs/manifests/config.pp:184 > > katello-0.1.135-2.el6.x86_64 > katello-all-0.1.135-2.el6.x86_64 > katello-certs-tools-1.0.1-2.el6.noarch > katello-cli-0.1.23-3.el6.noarch > katello-cli-common-0.1.23-3.el6.noarch > katello-common-0.1.135-2.el6.noarch > katello-configure-0.1.31-2.el6.noarch > katello-glue-candlepin-0.1.135-2.el6.x86_64 > katello-glue-foreman-0.1.135-2.el6.x86_64 > katello-glue-pulp-0.1.135-2.el6.x86_64 > katello-qpid-broker-key-pair-1.0-1.noarch > katello-trusted-ssl-cert-1.0-1.noarch > > Mike followup, this only seems to occur on EL 6.2 but not on Fedora or or EL 6.1 or before. most likely caused by a difference in nss-tools: EL6.1: nss-tools-3.12.9-9.el6.x86_64 EL6.2: nss-tools-3.12.10-16.el6.x86_64 will reply here if I get more details. Mike From lzap at redhat.com Fri Dec 9 10:51:31 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 9 Dec 2011 11:51:31 +0100 Subject: [katello-devel] errors with latest packages In-Reply-To: <4EE1938D.8050808@redhat.com> References: <4EE160D5.6070702@redhat.com> <4EE1938D.8050808@redhat.com> Message-ID: <20111209105131.GD2951@lzapx.brq.redhat.com> Mike, the nss-datatase is unused currently, we can just remove this step. The easiest would be to replace cmd with something that just return true: cmd => "/bin/true" Done. On Thu, Dec 08, 2011 at 08:50:21PM -0800, Michael McCune wrote: > On 12/08/2011 05:13 PM, Mike McCune wrote: > ># katello-configure > >Starting Katello configuration > >The top-level log file is > >[/var/log/katello/katello-configure-20111209-091609/main.log] > >err: /Stage[main]/Certs::Config/Exec[create-nss-db]/returns: change from > >notrun to 0 failed: /bin/rm -f /etc/pki/katello/nssdb//*; certutil -N -d > >'/etc/pki/katello/nssdb/' -f '/etc/katello/nss_db_password-file'; > >certutil -A -d '/etc/pki/katello/nssdb/' -n 'ca' -t 'TCu,Cu,Tuw' -a -i > >'/usr/share/katello/KATELLO-TRUSTED-SSL-CERT'; certutil -A -d > >'/etc/pki/katello/nssdb/' -n 'broker' -t ',,' -a -i > >'/etc/pki/tls/certs/qpid-broker.crt'; certutil -A -d > >'/etc/pki/katello/nssdb/' -n 'tomcat' -t ',,' -a -i > >'/etc/pki/tls/certs/httpd-ssl.crt' returned 255 instead of one of [0] at > >/usr/share/katello/install/puppet/modules/certs/manifests/config.pp:184 > > > >katello-0.1.135-2.el6.x86_64 > >katello-all-0.1.135-2.el6.x86_64 > >katello-certs-tools-1.0.1-2.el6.noarch > >katello-cli-0.1.23-3.el6.noarch > >katello-cli-common-0.1.23-3.el6.noarch > >katello-common-0.1.135-2.el6.noarch > >katello-configure-0.1.31-2.el6.noarch > >katello-glue-candlepin-0.1.135-2.el6.x86_64 > >katello-glue-foreman-0.1.135-2.el6.x86_64 > >katello-glue-pulp-0.1.135-2.el6.x86_64 > >katello-qpid-broker-key-pair-1.0-1.noarch > >katello-trusted-ssl-cert-1.0-1.noarch > > > >Mike > > followup, this only seems to occur on EL 6.2 but not on Fedora or or > EL 6.1 or before. most likely caused by a difference in nss-tools: > > EL6.1: > > nss-tools-3.12.9-9.el6.x86_64 > > EL6.2: > > nss-tools-3.12.10-16.el6.x86_64 > > will reply here if I get more details. > > Mike > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Fri Dec 9 11:13:31 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 9 Dec 2011 12:13:31 +0100 Subject: [katello-devel] Non-ar search discussion In-Reply-To: <4EE0ED56.6040203@redhat.com> References: <4EE0ED56.6040203@redhat.com> Message-ID: <20111209111331.GE2951@lzapx.brq.redhat.com> As a long-time Apache Lucene user, advocate and contributor I am all in, but let's focus more on other features right now. I am okay with collecting requirements and open discussion. But I vote for giving this lower priority. ;-) Instead of full-blown indexing server, I would prefer Apache Lucene compatible, lightweight, custom-made solution. From the Satellite experience, we would be able to leverage just 2 % of what these servers offer. Comparing with packaging efforts and memory consumptions, I would rather avoid them. Most of the features are web crawling and ability to drop document of any type (Word, OpenOffice, PDF) there and we don't need such features. We need to index our custom stuff like packages or other records. IMHO Lucene compatible API is so consistent, clean and easy to understand. LZ On Thu, Dec 08, 2011 at 12:01:10PM -0500, Justin Sherrill wrote: > Hi All, > > I've been tasked with looking into a mechanism to do non-ar search. > While amos had leaned towards a search proxy, I'm more inclined to > utilize a full text search server. I've put some of the pros/cons > here: > > https://fedorahosted.org/katello/wiki/KatelloSearchRevamp > > Along with a list of things that need to be searched. > > Thoughts? Comments? > > -Justin > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From jsherril at redhat.com Fri Dec 9 14:23:01 2011 From: jsherril at redhat.com (Justin Sherrill) Date: Fri, 09 Dec 2011 09:23:01 -0500 Subject: [katello-devel] Non-ar search discussion In-Reply-To: <20111209111331.GE2951@lzapx.brq.redhat.com> References: <4EE0ED56.6040203@redhat.com> <20111209111331.GE2951@lzapx.brq.redhat.com> Message-ID: <4EE219C5.3000305@redhat.com> On 12/09/2011 06:13 AM, Lukas Zapletal wrote: > As a long-time Apache Lucene user, advocate and contributor I am all in, > but let's focus more on other features right now. I am okay with > collecting requirements and open discussion. But I vote for giving this > lower priority. ;-) I don't think package/errata searching is a low priority for V1 :) > Instead of full-blown indexing server, I would prefer Apache Lucene > compatible, lightweight, custom-made solution. From the Satellite > experience, we would be able to leverage just 2 % of what these servers > offer. Comparing with packaging efforts and memory consumptions, I would > rather avoid them. Most of the features are web crawling and ability to > drop document of any type (Word, OpenOffice, PDF) there and we don't > need such features. We need to index our custom stuff like packages or > other records. > > IMHO Lucene compatible API is so consistent, clean and easy to understand. Have you looked at elastic search? Its the one I'm currently favouring and the one that delta cloud was seriously considering (they ended up pushing full text search for v2). It is very lightweight and simple and from a quick glance all its dependencies are already in fedora (although not all at the correct version). -Justin > LZ > > On Thu, Dec 08, 2011 at 12:01:10PM -0500, Justin Sherrill wrote: >> Hi All, >> >> I've been tasked with looking into a mechanism to do non-ar search. >> While amos had leaned towards a search proxy, I'm more inclined to >> utilize a full text search server. I've put some of the pros/cons >> here: >> >> https://fedorahosted.org/katello/wiki/KatelloSearchRevamp >> >> Along with a list of things that need to be searched. >> >> Thoughts? Comments? >> >> -Justin >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel From inecas at redhat.com Fri Dec 9 14:35:27 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Fri, 09 Dec 2011 15:35:27 +0100 Subject: [katello-devel] Non-ar search discussion In-Reply-To: <4EE219C5.3000305@redhat.com> References: <4EE0ED56.6040203@redhat.com> <20111209111331.GE2951@lzapx.brq.redhat.com> <4EE219C5.3000305@redhat.com> Message-ID: <4EE21CAF.5060102@redhat.com> On 12/09/2011 03:23 PM, Justin Sherrill wrote: > Have you looked at elastic search? Its the one I'm currently > favouring and the one that delta cloud was seriously considering (they > ended up pushing full text search for v2). It is very lightweight > and simple and from a quick glance all its dependencies are already in > fedora (although not all at the correct version). > > -Justin Just a note, a Ruby client [1] for the ES is maintained by a Czech developer. For Lukas - ES is based on Lucene, with nice RESTful Interface. [1] - https://github.com/karmi/tire From lzap at redhat.com Fri Dec 9 15:08:17 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 9 Dec 2011 16:08:17 +0100 Subject: [katello-devel] Missing progress bars for katello-configure Message-ID: <20111209150816.GJ2951@lzapx.brq.redhat.com> Hey, I am missing those nice progress bars. Do you see the same behavior? Thanks -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From jweiss at redhat.com Fri Dec 9 16:30:49 2011 From: jweiss at redhat.com (Jeff Weiss) Date: Fri, 09 Dec 2011 11:30:49 -0500 Subject: [katello-devel] Missing progress bars for katello-configure In-Reply-To: <20111209150816.GJ2951@lzapx.brq.redhat.com> References: <20111209150816.GJ2951@lzapx.brq.redhat.com> Message-ID: <1323448249.25307.8.camel@jweiss> On Fri, 2011-12-09 at 16:08 +0100, Lukas Zapletal wrote: > Hey, > > I am missing those nice progress bars. Do you see the same behavior? > > Thanks > For a while I saw them intermittently (one install would show them, then another would not). Now I don't see them at all. -jeff From thomasmckay at redhat.com Mon Dec 12 18:39:00 2011 From: thomasmckay at redhat.com (Tom McKay) Date: Mon, 12 Dec 2011 13:39:00 -0500 (EST) Subject: [katello-devel] subscriptions displayed in katello/headpin In-Reply-To: Message-ID: <6516ee35-2b3f-4319-aa90-00cbe53ff891@zmail16.collab.prod.int.phx2.redhat.com> There are four places that subscriptions are displayed in the UI. Each of these locations serve a specific purpose for the user, or is supposed to. Below is described how subscription info is displayed now, with suggestions/questions for how it should really work. 1. CONTENT MANAGEMENT -> RED HAT The purpose of this page is to: Display all the subscriptions available to an organization that were provided by importing a Red Hat manifest. * Subscriptions are displayed that were made available by "Upload Subscription Manifest" * Subscriptions w/ the same stacking id are grouped together * Bonus pool subscriptions are grouped under subscription that generated it (BZ 766694) ? Should bonus pools be listed here at all? ? Bonus pools, because they are available only to guests on a specific host, need host name info displayed. 2. SYSTEMS -> SUBSCRIPTIONS The purpose of this page is to: Show subscriptions available to a system and allow them to be subscribed to and unsubscribed from. * Only subscriptions available to the system are displayed. * A subscription's installed products are listed under the subscription. * Subscriptions are not grouped in any way. ? Confirm that virtual-only subscriptions are not displayed for physical machines. ? Confirm that bonus pool subscriptions are displayed only for guests of the host that generated them. ? Bonus pools, because they are available only to guests on a specific host, need host name info displayed. 3. SYSTEMS -> ACTIVATION KEYS The purpose of this page is to: Group subscriptions so that they may be easily subscribed to by a system. * All subscriptions available to an organization are displayed. * Subscriptions are grouped by stacking id. (Parent expands to show all w/ same stacking id.) * Subscriptions are grouped by bonus pool. (Parent expands to show generated bonus pool subscriptions.) ? Bonus pools, because they are available only to guests on a specific host, need host name info displayed. Example: + subscription-for-marketing-product-A -- subscription-for-marketing-product-A stacking-id=1 -- subscription-for-marketing-product-B stacking-id=1 -- subscription-for-marketing-product-A bonus-pool=true host=1.1.1.1 -- subscription-for-marketing-product-A bonus-pool=true host=1.1.1.2 -- subscription-for-marketing-product-A bonus-pool=true host=1.1.1.3 4. ORGANIZATIONS -> SUBSCRIPTIONS The purpose of this page is to: Display all the subscriptions available to an organization. * Subscriptions w/ the same stacking id are grouped together * Bonus pool subscriptions are grouped under subscription that generated it (BZ 766694) ? Bonus pools, because they are available only to guests on a specific host, need host name info displayed. From jrist at redhat.com Mon Dec 12 23:34:30 2011 From: jrist at redhat.com (Jason Rist) Date: Mon, 12 Dec 2011 16:34:30 -0700 Subject: [katello-devel] Merged Dashboard Popout Message-ID: <4EE68F86.6010507@redhat.com> Merged in dashboard popouts. All spec tests passed and it looked good. You will have to do a `rake db:migrate`. Thanks, Jason -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From lzap at redhat.com Tue Dec 13 10:09:44 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 13 Dec 2011 11:09:44 +0100 Subject: [katello-devel] Please keep seeds.rb script idempotent Message-ID: <20111213100944.GB3159@lzapx.brq.redhat.com> Hello, I rewrote our seeds.rb script so it's idempotent again. What it means? It means we are able to run it multiple times without any bite. Instead of creating all users, roles and providers blindly, it first try to find it in the Katello database and if it finds it - it does not create it. It's that simple. The rule is to maintain katello and backend databases in sync. If you really want the seeds script to re-create everything again, delete whole Katello database AND all backend engines databases as well. If you only do the former, it will likely fail as our orchestration code won't be able to make the changes in backend engines. >From the technical perspective, instead of calling create! we must use find_or_create_by_something. But please note since those find_or_create methods does not have exclamation versions, we need to explicitly check the result against nil (and it's good to check errors array as well). I rewrote all the stuff so it's working now. I added a warning comment on the top of the file. Please stick with my recommendation as this is the very first step to support of upgrades. ps - I am also working on having Puppet idempotent. So far it looks good, at the end of the day we should be able to re-execute katello-configure without any harm. But as of today, THIS IS NOT SUPPORTED YET. Due to various issues (cpsetup, uknown race conditions) it will likely fail. Note to QA: Please everytime you report an issue with the installer, add a note if this is a first run of katello-configure on a fresh machine. Thanks. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Tue Dec 13 13:15:57 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 13 Dec 2011 14:15:57 +0100 Subject: [katello-devel] Katello packages noarch again Message-ID: <20111213131557.GE3159@lzapx.brq.redhat.com> Hello, with Katello/Headpin package breakup we forgot to mark all packages as noarch. Only katello-common had the noarch statement, others were x86-64. I added the statements back in our SPEC file. Just to let you know, I hope that won't break something... -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Tue Dec 13 15:58:38 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 13 Dec 2011 16:58:38 +0100 Subject: [katello-devel] Easy profiling setup Message-ID: <20111213155838.GH3159@lzapx.brq.redhat.com> Hey, was curious why Ruby takes so much CPU while importing a big manifest. I tried this very easy approach to see some profiling information: rails profiler 'User.current = User.find(1); \ Provider.find(1).import_manifest "/root/manifest.zip"' And it did not point out on a particular flaw in our codebase. It's just slow. https://fedorahosted.org/katello/wiki/Profiling But the command could be useful for you, so sharing. LZ -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From mmccune at redhat.com Tue Dec 13 17:04:39 2011 From: mmccune at redhat.com (Mike McCune) Date: Tue, 13 Dec 2011 09:04:39 -0800 Subject: [katello-devel] Easy profiling setup In-Reply-To: <20111213155838.GH3159@lzapx.brq.redhat.com> References: <20111213155838.GH3159@lzapx.brq.redhat.com> Message-ID: <4EE785A7.5020006@redhat.com> On 12/13/2011 07:58 AM, Lukas Zapletal wrote: > Hey, > > was curious why Ruby takes so much CPU while importing a big manifest. I > tried this very easy approach to see some profiling information: > > rails profiler 'User.current = User.find(1); \ > Provider.find(1).import_manifest "/root/manifest.zip"' > > And it did not point out on a particular flaw in our codebase. It's just > slow. heh, I think there is a lot of overhead with hundreds of PUTs to pulp to create the repos. If you tail /var/log/httpd/ssl_kt_access_log you will note that the vast majority of the time spent during manifest import is looping over the repos and PUTing them into Pulp's DB. > > https://fedorahosted.org/katello/wiki/Profiling > > But the command could be useful for you, so sharing. thanks, will definitely use it. Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From paji at redhat.com Tue Dec 13 18:30:19 2011 From: paji at redhat.com (Partha Aji) Date: Tue, 13 Dec 2011 13:30:19 -0500 (EST) Subject: [katello-devel] Easy profiling setup In-Reply-To: <4EE785A7.5020006@redhat.com> Message-ID: There was an idea proposed by Mike a while back on only creating repos in pulp if they were 'enabled'. So import manifest for example would import things to candlepin but not create repos in pulp. The user would go to the red hat providers page and then enable repos specifically, there by creating the repo. We originally shot down the idea by saying there might be information in the manifest that might be available to us while creating pulp repos on manifest import (as in we'd lose this information if we created a repo later on enablement). But I haven't confirmed that this is the case. May be we'd not need extra information to create pulp repos and thus create em on enablement. If we can prevent 400+ pulp repo create requests (out of which only 5 /6 might be enabled and thus usable anyway) and do it selectively using the enablement criteria, manifest import performance will be magnitudes faster. Partha ----- Original Message ----- > From: "Mike McCune" > To: katello-devel at redhat.com > Sent: Tuesday, December 13, 2011 12:04:39 PM > Subject: Re: [katello-devel] Easy profiling setup > > On 12/13/2011 07:58 AM, Lukas Zapletal wrote: > > Hey, > > > > was curious why Ruby takes so much CPU while importing a big > > manifest. I > > tried this very easy approach to see some profiling information: > > > > rails profiler 'User.current = User.find(1); \ > > Provider.find(1).import_manifest "/root/manifest.zip"' > > > > And it did not point out on a particular flaw in our codebase. It's > > just > > slow. > > heh, I think there is a lot of overhead with hundreds of PUTs to pulp > to > create the repos. If you tail /var/log/httpd/ssl_kt_access_log you > will > note that the vast majority of the time spent during manifest import > is > looping over the repos and PUTing them into Pulp's DB. > > > > > https://fedorahosted.org/katello/wiki/Profiling > > > > But the command could be useful for you, so sharing. > > thanks, will definitely use it. > > Mike > -- > Mike McCune > mmccune AT redhat.com > Red Hat Engineering | Portland, OR > Systems Management | 650.254.4248 > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel > From lzap at redhat.com Wed Dec 14 08:49:10 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Wed, 14 Dec 2011 09:49:10 +0100 Subject: [katello-devel] Easy profiling setup In-Reply-To: References: <4EE785A7.5020006@redhat.com> Message-ID: <20111214084910.GA2447@lzapx.brq.redhat.com> This is a good idea (importing only enabled). Yeah I know it is accessing Pulp. I was just wondering why Thin consumes 60 % CPU time while Pulp consumes roughly 5 % CPU. I would expect Thin to be "waiting" for responses. From my top it seems Thin is doing some hard work, maybe harder than Pulp. We are doing some string replacements and other things - so I wanted to do a quick check-up what's going on. And the outcome is we do not have a major flaw in our codebase regarding manifest imports. At least it seems there is not any. Not talking about memory consumption which is higher than Candlepin (a Java stack) during import. But this is another story, I was only profiling method calls. LZ On Tue, Dec 13, 2011 at 01:30:19PM -0500, Partha Aji wrote: > > There was an idea proposed by Mike a while back on only creating repos in pulp if they were 'enabled'. > So import manifest for example would import things to candlepin but not create repos in pulp. The user > would go to the red hat providers page and then enable repos specifically, there by creating the repo. > > We originally shot down the idea by saying there might be information in the manifest that might be > available to us while creating pulp repos on manifest import (as in we'd lose this information > if we created a repo later on enablement). But I haven't confirmed that this is the case. > May be we'd not need extra information to create pulp repos and thus create em on enablement. > > If we can prevent 400+ pulp repo create requests (out of which only 5 /6 might be enabled and thus usable anyway) > and do it selectively using the enablement criteria, manifest import performance will be magnitudes faster. > > > Partha > > ----- Original Message ----- > > From: "Mike McCune" > > To: katello-devel at redhat.com > > Sent: Tuesday, December 13, 2011 12:04:39 PM > > Subject: Re: [katello-devel] Easy profiling setup > > > > On 12/13/2011 07:58 AM, Lukas Zapletal wrote: > > > Hey, > > > > > > was curious why Ruby takes so much CPU while importing a big > > > manifest. I > > > tried this very easy approach to see some profiling information: > > > > > > rails profiler 'User.current = User.find(1); \ > > > Provider.find(1).import_manifest "/root/manifest.zip"' > > > > > > And it did not point out on a particular flaw in our codebase. It's > > > just > > > slow. > > > > heh, I think there is a lot of overhead with hundreds of PUTs to pulp > > to > > create the repos. If you tail /var/log/httpd/ssl_kt_access_log you > > will > > note that the vast majority of the time spent during manifest import > > is > > looping over the repos and PUTing them into Pulp's DB. > > > > > > > > https://fedorahosted.org/katello/wiki/Profiling > > > > > > But the command could be useful for you, so sharing. > > > > thanks, will definitely use it. > > > > Mike > > -- > > Mike McCune > > mmccune AT redhat.com > > Red Hat Engineering | Portland, OR > > Systems Management | 650.254.4248 > > > > _______________________________________________ > > katello-devel mailing list > > katello-devel at redhat.com > > https://www.redhat.com/mailman/listinfo/katello-devel > > > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From inecas at redhat.com Wed Dec 14 13:13:55 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Wed, 14 Dec 2011 14:13:55 +0100 Subject: [katello-devel] Non-ar search discussion In-Reply-To: <4EE21CAF.5060102@redhat.com> References: <4EE0ED56.6040203@redhat.com> <20111209111331.GE2951@lzapx.brq.redhat.com> <4EE219C5.3000305@redhat.com> <4EE21CAF.5060102@redhat.com> Message-ID: <4EE8A113.1000909@redhat.com> Btw. regarding the %subject% there are recently 3 episodes on RailsCasts that might be helpful. http://railscasts.com/episodes/306-elasticsearch-part-1 http://railscasts.com/episodes/307-elasticsearch-part-2 http://railscasts.com/episodes/120-thinking-sphinx-revised -- Ivan On 12/09/2011 03:35 PM, Ivan Ne?as wrote: > On 12/09/2011 03:23 PM, Justin Sherrill wrote: >> Have you looked at elastic search? Its the one I'm currently >> favouring and the one that delta cloud was seriously considering >> (they ended up pushing full text search for v2). It is very >> lightweight and simple and from a quick glance all its dependencies >> are already in fedora (although not all at the correct version). >> >> -Justin > > Just a note, a Ruby client [1] for the ES is maintained by a Czech > developer. For Lukas - ES is based on Lucene, with nice RESTful > Interface. > > [1] - https://github.com/karmi/tire > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From dgoodwin at rm-rf.ca Wed Dec 14 15:30:03 2011 From: dgoodwin at rm-rf.ca (Devan Goodwin) Date: Wed, 14 Dec 2011 11:30:03 -0400 Subject: [katello-devel] Per-Arch Promotion Message-ID: Ivan and I are wondering, if you get a content set in a manifest with a URL like: /content/awesomeos/server/10/$releasever/$basearch/os And assuming there is some mechanism by which we figure out which $releasever's and $basearch's there are so we know what to sync (I am fuzzy on this but I believe there is a plan), do we need to support "per-arch" promotion to an environment? i.e. I only want to promote x86_64 and not i386 to an environment. Or can we assume that this content is promoted as a logical unit, all arches and released versions. Thanks, Devan From inecas at redhat.com Wed Dec 14 15:36:41 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Wed, 14 Dec 2011 16:36:41 +0100 Subject: [katello-devel] Per-Arch Promotion In-Reply-To: References: Message-ID: <4EE8C289.7080507@redhat.com> On 12/14/2011 04:30 PM, Devan Goodwin wrote: > Ivan and I are wondering, if you get a content set in a manifest with > a URL like: > > /content/awesomeos/server/10/$releasever/$basearch/os > > And assuming there is some mechanism by which we figure out which > $releasever's and $basearch's there are so we know what to sync (I am > fuzzy on this but I believe there is a plan), do we need to support > "per-arch" promotion to an environment? Just a note: The mechanism right now is repositories black-list. > > i.e. I only want to promote x86_64 and not i386 to an environment. > > Or can we assume that this content is promoted as a logical unit, all > arches and released versions. > > Thanks, > > Devan > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From mmccune at redhat.com Thu Dec 15 00:02:36 2011 From: mmccune at redhat.com (Mike McCune) Date: Wed, 14 Dec 2011 16:02:36 -0800 Subject: [katello-devel] Inbound patch to compress assets Message-ID: <4EE9391C.1030003@redhat.com> Adding changes to ensure we compress our assets served up by Katello. fedorahosted.org is in outage but when it is back I'll be pushing this and rebuilding RPMs to test some more. So far I've not noticed any downsides. commit 95cee10b749c36ee5fb264b5cee3cb4d4120b43c Author: Mike McCune Date: Wed Dec 14 14:18:38 2011 -0800 767812 - compress our javascript and CSS diff --git a/puppet/modules/katello/templates/etc/httpd/conf.d/katello.conf.erb b/puppet/modules/katello/templates/etc/httpd/conf.d/katello.conf.erb index 9d1be6f..4db9345 100644 --- a/puppet/modules/katello/templates/etc/httpd/conf.d/katello.conf.erb +++ b/puppet/modules/katello/templates/etc/httpd/conf.d/katello.conf.erb @@ -13,6 +13,9 @@ NameVirtualHost *:443 ProxyPreserveHost Off RequestHeader set X_FORWARDED_PROTO 'https' + # Performance related configurations + AddOutputFilterByType DEFLATE text/html text/css application/x-javascript application/javascript + Timeout 5400 ProxyTimeout 5400 diff --git a/src/config/assets.yml b/src/config/assets.yml index 59284f7..86bfd58 100644 --- a/src/config/assets.yml +++ b/src/config/assets.yml @@ -1,5 +1,7 @@ embed_assets: on javascript_compressor: yui +compress_assets: on +gzip_assets: on javascripts: common: diff --git a/src/katello.spec b/src/katello.spec index 9bd4e40..53899a4 100644 --- a/src/katello.spec +++ b/src/katello.spec @@ -159,7 +159,8 @@ compass compile #generate Rails JS/CSS/... assets echo Generating Rails assets... -jammit +jammit --config config/assets.yml -f + -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From mmccune at redhat.com Thu Dec 15 01:03:30 2011 From: mmccune at redhat.com (Mike McCune) Date: Wed, 14 Dec 2011 17:03:30 -0800 Subject: [katello-devel] expect bugzilla churn Message-ID: <4EE94762.7000206@redhat.com> I'm going through, triaging, re-assigning, commenting and mucking with our open bugs. just a heads up! Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From lzap at redhat.com Thu Dec 15 08:31:22 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 15 Dec 2011 09:31:22 +0100 Subject: [katello-devel] Fedorahosted down Message-ID: <20111215083122.GA5603@lzapx.brq.redhat.com> Hey, fedorahosted.org is down. That means: - git - wiki FYI -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Thu Dec 15 08:45:01 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 15 Dec 2011 09:45:01 +0100 Subject: [katello-devel] Having a default environment Message-ID: <20111215084501.GC5603@lzapx.brq.redhat.com> Hey, so to create or not to create a default environment. I think we already discussed that and the outcome was not to create one. https://bugzilla.redhat.com/show_bug.cgi?id=767665 Thanks -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From mmccune at redhat.com Thu Dec 15 19:47:55 2011 From: mmccune at redhat.com (Mike McCune) Date: Thu, 15 Dec 2011 11:47:55 -0800 Subject: [katello-devel] Fedorahosted down In-Reply-To: <20111215083122.GA5603@lzapx.brq.redhat.com> References: <20111215083122.GA5603@lzapx.brq.redhat.com> Message-ID: <4EEA4EEB.5050404@redhat.com> On 12/15/2011 12:31 AM, Lukas Zapletal wrote: > Hey, > > fedorahosted.org is down. That means: > > - git > - wiki > > FYI > it is back up but they switched IPs so you may have to add this to your /etc/hosts: 66.135.62.191 fedorahosted.org git.fedorahosted.org hg.fedorahosted.org svn.fedorahosted.org lists.fedorahosted.org bzr.fedorahosted.org -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From jrist at redhat.com Thu Dec 15 20:36:09 2011 From: jrist at redhat.com (Jason Rist) Date: Thu, 15 Dec 2011 13:36:09 -0700 Subject: [katello-devel] Fedorahosted down In-Reply-To: <4EEA4EEB.5050404@redhat.com> References: <20111215083122.GA5603@lzapx.brq.redhat.com> <4EEA4EEB.5050404@redhat.com> Message-ID: <4EEA5A39.7050209@redhat.com> On 12/15/2011 12:47 PM, Mike McCune wrote: > On 12/15/2011 12:31 AM, Lukas Zapletal wrote: >> Hey, >> >> fedorahosted.org is down. That means: >> >> - git >> - wiki >> >> FYI >> > > it is back up but they switched IPs so you may have to add this to your > /etc/hosts: > > > 66.135.62.191 fedorahosted.org git.fedorahosted.org > hg.fedorahosted.org svn.fedorahosted.org lists.fedorahosted.org > bzr.fedorahosted.org > beav also pointed out restarting nscd or doing `nscd -i hosts` -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From lzap at redhat.com Fri Dec 16 10:10:16 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 16 Dec 2011 11:10:16 +0100 Subject: [katello-devel] Unable to set quantity in the activation keys subscriptions Message-ID: <20111216101016.GA6981@lzapx.brq.redhat.com> Hello, today I wanted to set quantity in the activation keys screen for a subscriptions and I was not able to. I know there was some refactoring. But we still have an "allocated" column in the database. How do I set it? How is this expected to work during activation key registration? I think Bryan mentioned to consume number of processors (from the facts). Should we get rid of this column then? Thanks -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From inecas at redhat.com Fri Dec 16 12:43:57 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Fri, 16 Dec 2011 13:43:57 +0100 Subject: [katello-devel] Unable to set quantity in the activation keys subscriptions In-Reply-To: <20111216101016.GA6981@lzapx.brq.redhat.com> References: <20111216101016.GA6981@lzapx.brq.redhat.com> Message-ID: <4EEB3D0D.10801@redhat.com> If I recall it correctly, (correct me, if I'm wrong) when subscribing directly (without activation keys), CP can figure out how many subscriptions the system needs (from number of socekts etc.). When using activation keys, we duplicate this functionality, and so we need to implement this logic (or get that info from CP?) into Katello as well. Therefore the number of allocations is not valid any more I think and it should go away. -- Ivan On 12/16/2011 11:10 AM, Lukas Zapletal wrote: > Hello, > > today I wanted to set quantity in the activation keys screen for a > subscriptions and I was not able to. > > I know there was some refactoring. But we still have an "allocated" > column in the database. How do I set it? How is this expected to work > during activation key registration? > > I think Bryan mentioned to consume number of processors (from the > facts). Should we get rid of this column then? > > Thanks > -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From bkearney at redhat.com Fri Dec 16 12:59:20 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Fri, 16 Dec 2011 07:59:20 -0500 Subject: [katello-devel] Unable to set quantity in the activation keys subscriptions In-Reply-To: <4EEB3D0D.10801@redhat.com> References: <20111216101016.GA6981@lzapx.brq.redhat.com> <4EEB3D0D.10801@redhat.com> Message-ID: <4EEB40A8.2040401@redhat.com> On 12/16/2011 07:43 AM, Ivan Ne?as wrote: > If I recall it correctly, (correct me, if I'm wrong) when subscribing > directly (without activation keys), CP can figure out how many > subscriptions the system needs (from number of socekts etc.). > > When using activation keys, we duplicate this functionality, and so we > need to implement this logic (or get that info from CP?) into Katello as > well. Therefore the number of allocations is not valid any more I think > and it should go away. This is probably the better way to go. -- bk From jrist at redhat.com Fri Dec 16 14:31:00 2011 From: jrist at redhat.com (Jason Rist) Date: Fri, 16 Dec 2011 07:31:00 -0700 Subject: [katello-devel] Unable to set quantity in the activation keys subscriptions In-Reply-To: <4EEB3D0D.10801@redhat.com> References: <20111216101016.GA6981@lzapx.brq.redhat.com> <4EEB3D0D.10801@redhat.com> Message-ID: <4EEB5624.4000804@redhat.com> On 12/16/2011 05:43 AM, Ivan Ne?as wrote: > If I recall it correctly, (correct me, if I'm wrong) when subscribing > directly (without activation keys), CP can figure out how many > subscriptions the system needs (from number of socekts etc.). > > When using activation keys, we duplicate this functionality, and so we > need to implement this logic (or get that info from CP?) into Katello as > well. Therefore the number of allocations is not valid any more I think > and it should go away. > > -- Ivan > > On 12/16/2011 11:10 AM, Lukas Zapletal wrote: >> Hello, >> >> today I wanted to set quantity in the activation keys screen for a >> subscriptions and I was not able to. >> >> I know there was some refactoring. But we still have an "allocated" >> column in the database. How do I set it? How is this expected to work >> during activation key registration? >> >> I think Bryan mentioned to consume number of processors (from the >> facts). Should we get rid of this column then? >> >> Thanks >> > > Yes, we removed it from the UI for the above reasons. -J -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From lzap at redhat.com Fri Dec 16 14:32:23 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 16 Dec 2011 15:32:23 +0100 Subject: [katello-devel] Unable to set quantity in the activation keys subscriptions In-Reply-To: <4EEB40A8.2040401@redhat.com> References: <20111216101016.GA6981@lzapx.brq.redhat.com> <4EEB3D0D.10801@redhat.com> <4EEB40A8.2040401@redhat.com> Message-ID: <20111216143223.GC6981@lzapx.brq.redhat.com> So assuming this (please confirm): a) Will not use allocated column. Will remove it from our codebase. b) Will use number of sockets as the allocation number. c) For multiple products I will threat all the pools as one. d) For multiple products, I will make the same allocation for all products. e) For multiple products, I will consume from the pool which starts most recently. LZ On Fri, Dec 16, 2011 at 07:59:20AM -0500, Bryan Kearney wrote: > On 12/16/2011 07:43 AM, Ivan Ne?as wrote: > >If I recall it correctly, (correct me, if I'm wrong) when subscribing > >directly (without activation keys), CP can figure out how many > >subscriptions the system needs (from number of socekts etc.). > > > >When using activation keys, we duplicate this functionality, and so we > >need to implement this logic (or get that info from CP?) into Katello as > >well. Therefore the number of allocations is not valid any more I think > >and it should go away. > > > This is probably the better way to go. > > -- bk > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From jmatthew at redhat.com Fri Dec 16 14:54:11 2011 From: jmatthew at redhat.com (John Matthews) Date: Fri, 16 Dec 2011 09:54:11 -0500 (EST) Subject: [katello-devel] Pulp's changed /etc/pki/content to /etc/pki/pulp/content In-Reply-To: Message-ID: <5a5776eb-b0ad-46c5-8f1f-37e36afb41ab@zmail13.collab.prod.int.phx2.redhat.com> Pulp has changed where it stores certs. This may have an impact to Candlepin and the location of the CRL it creates when integrated with Pulp. The pulp-migrate script will handle upgrades. It will move cert files from the old location to /etc/pki/pulp/contentXXXX If you see any problems please let me know. From inecas at redhat.com Fri Dec 16 15:08:38 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Fri, 16 Dec 2011 16:08:38 +0100 Subject: [katello-devel] Pulp's changed /etc/pki/content to /etc/pki/pulp/content In-Reply-To: <5a5776eb-b0ad-46c5-8f1f-37e36afb41ab@zmail13.collab.prod.int.phx2.redhat.com> References: <5a5776eb-b0ad-46c5-8f1f-37e36afb41ab@zmail13.collab.prod.int.phx2.redhat.com> Message-ID: <4EEB5EF6.60203@redhat.com> Hi We configure pulp.conf and repo_auth.conf explicitly, including crl and cert location, so it should not affect us at all. We may change the installer to reflect the change once it's in CR? Is it already there? -- Ivan On 12/16/2011 03:54 PM, John Matthews wrote: > Pulp has changed where it stores certs. > This may have an impact to Candlepin and the location of the CRL it creates when integrated with Pulp. > > The pulp-migrate script will handle upgrades. It will move cert files from the old location to /etc/pki/pulp/contentXXXX > > If you see any problems please let me know. > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From jmatthew at redhat.com Fri Dec 16 15:14:11 2011 From: jmatthew at redhat.com (John Matthews) Date: Fri, 16 Dec 2011 10:14:11 -0500 (EST) Subject: [katello-devel] Pulp's changed /etc/pki/content to /etc/pki/pulp/content In-Reply-To: <4EEB5EF6.60203@redhat.com> Message-ID: <8c5b269d-3beb-49c8-abdb-b44d193a0d6e@zmail13.collab.prod.int.phx2.redhat.com> Hi Ivan, This change will be in CR20, which will be released in mid Jan. To confirm, it is not in CR19 which was released yesterday. If you are manually changing this value, you may need to adjust the SELinux policy. If you need assistance I'd be happy to help. Pulp is expecting that the content certs it manages will be labeled as "system_u:object_r:pulp_cert_t" We use the below rule to handle this: /etc/pki/pulp(/.*)? gen_context(system_u:object_r:pulp_cert_t,s0) ----- Original Message ----- > Hi > > We configure pulp.conf and repo_auth.conf explicitly, including crl > and > cert location, so it should not affect us at all. We may change the > installer to reflect the change once it's in CR? Is it already there? > > -- Ivan > On 12/16/2011 03:54 PM, John Matthews wrote: > > Pulp has changed where it stores certs. > > This may have an impact to Candlepin and the location of the CRL it > > creates when integrated with Pulp. > > > > The pulp-migrate script will handle upgrades. It will move cert > > files from the old location to /etc/pki/pulp/contentXXXX > > > > If you see any problems please let me know. > > > > _______________________________________________ > > katello-devel mailing list > > katello-devel at redhat.com > > https://www.redhat.com/mailman/listinfo/katello-devel > > > -- > Regards, > > Ivan Necas > Red Hat Czech s.r.o. Brno > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel > From bbuckingham at redhat.com Fri Dec 16 15:16:00 2011 From: bbuckingham at redhat.com (Brad Buckingham) Date: Fri, 16 Dec 2011 10:16:00 -0500 Subject: [katello-devel] Unable to set quantity in the activation keys subscriptions In-Reply-To: <4EEB5624.4000804@redhat.com> References: <20111216101016.GA6981@lzapx.brq.redhat.com> <4EEB3D0D.10801@redhat.com> <4EEB5624.4000804@redhat.com> Message-ID: <4EEB60B0.9010402@redhat.com> On 12/16/2011 09:31 AM, Jason Rist wrote: > On 12/16/2011 05:43 AM, Ivan Ne?as wrote: >> If I recall it correctly, (correct me, if I'm wrong) when subscribing >> directly (without activation keys), CP can figure out how many >> subscriptions the system needs (from number of socekts etc.). >> >> When using activation keys, we duplicate this functionality, and so we >> need to implement this logic (or get that info from CP?) into Katello as >> well. Therefore the number of allocations is not valid any more I think >> and it should go away. >> >> -- Ivan >> >> On 12/16/2011 11:10 AM, Lukas Zapletal wrote: >>> Hello, >>> >>> today I wanted to set quantity in the activation keys screen for a >>> subscriptions and I was not able to. >>> >>> I know there was some refactoring. But we still have an "allocated" >>> column in the database. How do I set it? How is this expected to work >>> during activation key registration? >>> >>> I think Bryan mentioned to consume number of processors (from the >>> facts). Should we get rid of this column then? >>> >>> Thanks >>> >> > Yes, we removed it from the UI for the above reasons. > > -J > Jason is correct. It was removed from the UI a few sprints ago based on discussion that indicated that assigning a quantity in the UI was not very useful and that the better approach was to allow auto-subscribing to allocate the proper number during system at registration. At the time, we did not remove the column from the DB as there were references to it in backend and the plan was for it to be removed when backend/cli changes were made to support the auto-subscribe. It seems that now is that time. :) thanks, Brad From inecas at redhat.com Fri Dec 16 15:26:30 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Fri, 16 Dec 2011 16:26:30 +0100 Subject: [katello-devel] Pulp's changed /etc/pki/content to /etc/pki/pulp/content In-Reply-To: <8c5b269d-3beb-49c8-abdb-b44d193a0d6e@zmail13.collab.prod.int.phx2.redhat.com> References: <8c5b269d-3beb-49c8-abdb-b44d193a0d6e@zmail13.collab.prod.int.phx2.redhat.com> Message-ID: <4EEB6326.2090509@redhat.com> On 12/16/2011 04:14 PM, John Matthews wrote: > Hi Ivan, > > This change will be in CR20, which will be released in mid Jan. To confirm, it is not in CR19 which was released yesterday. > > If you are manually changing this value, you may need to adjust the SELinux policy. If you need assistance I'd be happy to help. > Pulp is expecting that the content certs it manages will be labeled as "system_u:object_r:pulp_cert_t" > > We use the below rule to handle this: > /etc/pki/pulp(/.*)? gen_context(system_u:object_r:pulp_cert_t,s0) It will be better to update our installer scripts to meet this new location. Maybe Katello's BZ is good place for this kind of announcements so that there is lower probability we miss it (triage should help) and we can track the change. -- Ivan > > > > ----- Original Message ----- >> Hi >> >> We configure pulp.conf and repo_auth.conf explicitly, including crl >> and >> cert location, so it should not affect us at all. We may change the >> installer to reflect the change once it's in CR? Is it already there? >> >> -- Ivan >> On 12/16/2011 03:54 PM, John Matthews wrote: >>> Pulp has changed where it stores certs. >>> This may have an impact to Candlepin and the location of the CRL it >>> creates when integrated with Pulp. >>> >>> The pulp-migrate script will handle upgrades. It will move cert >>> files from the old location to /etc/pki/pulp/contentXXXX >>> >>> If you see any problems please let me know. >>> >>> _______________________________________________ >>> katello-devel mailing list >>> katello-devel at redhat.com >>> https://www.redhat.com/mailman/listinfo/katello-devel >> >> -- >> Regards, >> >> Ivan Necas >> Red Hat Czech s.r.o. Brno >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel >> -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From jmatthew at redhat.com Fri Dec 16 15:51:59 2011 From: jmatthew at redhat.com (John Matthews) Date: Fri, 16 Dec 2011 10:51:59 -0500 (EST) Subject: [katello-devel] Pulp's changed /etc/pki/content to /etc/pki/pulp/content In-Reply-To: <4EEB6326.2090509@redhat.com> Message-ID: <9b8d061a-be0a-49d9-8b74-d849449e1933@zmail13.collab.prod.int.phx2.redhat.com> ----- Original Message ----- > On 12/16/2011 04:14 PM, John Matthews wrote: > > Hi Ivan, > > > > This change will be in CR20, which will be released in mid Jan. To > > confirm, it is not in CR19 which was released yesterday. > > > > If you are manually changing this value, you may need to adjust the > > SELinux policy. If you need assistance I'd be happy to help. > > Pulp is expecting that the content certs it manages will be labeled > > as "system_u:object_r:pulp_cert_t" > > > > We use the below rule to handle this: > > /etc/pki/pulp(/.*)? gen_context(system_u:object_r:pulp_cert_t,s0) > It will be better to update our installer scripts to meet this new > location. Maybe Katello's BZ is good place for this kind of > announcements so that there is lower probability we miss it (triage > should help) and we can track the change. > > -- Ivan > > Good suggestion Ivan, here is the BZ: https://bugzilla.redhat.com/show_bug.cgi?id=768420 Thanks, -John From bkearney at redhat.com Fri Dec 16 20:03:48 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Fri, 16 Dec 2011 15:03:48 -0500 Subject: [katello-devel] Backed patch verification please Message-ID: <4EEBA424.6060709@redhat.com> Can someone who has the backend test suite running try this patch for me? Zeus tested it live and it works there, I want to make sure there are no regressions. -- bk -------------- next part -------------- A non-text attachment was scrubbed... Name: 0001-767798-Pass-in-cp-consumer-instead-of-cp-user-for-sy.patch Type: text/x-patch Size: 806 bytes Desc: not available URL: From mmccune at redhat.com Fri Dec 16 20:03:28 2011 From: mmccune at redhat.com (Mike McCune) Date: Fri, 16 Dec 2011 12:03:28 -0800 Subject: [katello-devel] Fwd: [Pulp-list] Advanced Query Format Message-ID: <4EEBA410.8000603@redhat.com> As big users of Pulp and its APIs they are going to be doing work for V2 to greatly enhance searching, sorting and how to retrieve stuff from Pulp. Please comment on pulp-list if you have any comments. -------- Original Message -------- Subject: [Pulp-list] Advanced Query Format Date: Fri, 16 Dec 2011 14:54:55 -0500 From: Jay Dobies To: pulp-list at redhat.com https://fedorahosted.org/pulp/wiki/GCUnitAssociations#queryformat It looks like in v2 we're going to be supporting both GET URLs for simple queries (query in the parameters) and a POST URL for advanced search (letting us do more complicated AND/OR/IN/GT/LT types of operations). Had I started with a simpler area than unit associations, I'd be proposing a standard format to use for all advanced queries. But the unit associations queries are more complicated and will need some variations. The page above outlines my current thoughts on what that format looks like. I'd guess that once we settle on it, the standard query format will be a simplified version of this. I'm curious what people think. I'm open to new options, but I'd like to get started on this early next week, so the sooner you can propose them, the easier it'll be to change what I have written on that page. Thanks :D -- Jay Dobies Freenode: jdob @ #pulp http://pulpproject.org | http://blog.pulpproject.org _______________________________________________ Pulp-list mailing list Pulp-list at redhat.com https://www.redhat.com/mailman/listinfo/pulp-list From gkhachik at redhat.com Mon Dec 19 11:09:36 2011 From: gkhachik at redhat.com (Garik Khachikyan) Date: Mon, 19 Dec 2011 12:09:36 +0100 Subject: [katello-devel] Backed patch verification please In-Reply-To: <4EEBA424.6060709@redhat.com> References: <4EEBA424.6060709@redhat.com> Message-ID: <4EEF1B70.4030806@redhat.com> On 12/16/2011 09:03 PM, Bryan Kearney wrote: > Can someone who has the backend test suite running try this patch for > me? Zeus tested it live and it works there, I want to make sure there > are no regressions. > > -- bk > > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel Hi Bryan, I'll take the latest git (apply the patch) and make run of BE tests we have recently. Will keep you updated with reports then. Thanks, Garik. From gkhachik at redhat.com Mon Dec 19 11:58:33 2011 From: gkhachik at redhat.com (Garik Khachikyan) Date: Mon, 19 Dec 2011 12:58:33 +0100 Subject: [katello-devel] reponame structure changed for RHSM ? Message-ID: <4EEF26E9.2020104@redhat.com> Hi team, The recent tests detected that Katello reponame (which system gets subscribed to) seems to be changed: Like my repo named: repoBPM_pulpRHEL6_1324294815 became "22-repoBPM_pulpRHEL6_1324294815" for the yum repolist. (Repo is custom yum repo of Pulp RHEL6) So let me ask: where the "22-" appeared from ? Is it a regression bug or new feature? Thanks, Garik From inecas at redhat.com Mon Dec 19 12:22:52 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Mon, 19 Dec 2011 13:22:52 +0100 Subject: [katello-devel] reponame structure changed for RHSM ? In-Reply-To: <4EEF26E9.2020104@redhat.com> References: <4EEF26E9.2020104@redhat.com> Message-ID: <4EEF2C9C.5000705@redhat.com> On 12/19/2011 12:58 PM, Garik Khachikyan wrote: > Hi team, > > The recent tests detected that Katello reponame (which system gets > subscribed to) seems to be changed: > > Like my repo named: repoBPM_pulpRHEL6_1324294815 became > "22-repoBPM_pulpRHEL6_1324294815" for the yum repolist. > > (Repo is custom yum repo of Pulp RHEL6) > > So let me ask: where the "22-" appeared from ? Is it a regression bug > or new feature? It's a feature - fix for https://bugzilla.redhat.com/show_bug.cgi?id=758219, this number is there to keep repo id's in CP unique. -- Ivan > > > Thanks, > Garik > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From gkhachik at redhat.com Mon Dec 19 13:49:35 2011 From: gkhachik at redhat.com (Garik Khachikyan) Date: Mon, 19 Dec 2011 14:49:35 +0100 Subject: [katello-devel] Backed patch verification please In-Reply-To: <4EEF1B70.4030806@redhat.com> References: <4EEBA424.6060709@redhat.com> <4EEF1B70.4030806@redhat.com> Message-ID: <4EEF40EF.1010707@redhat.com> On 12/19/2011 12:09 PM, Garik Khachikyan wrote: > On 12/16/2011 09:03 PM, Bryan Kearney wrote: >> Can someone who has the backend test suite running try this patch for >> me? Zeus tested it live and it works there, I want to make sure there >> are no regressions. >> >> -- bk >> >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel > Hi Bryan, > > I'll take the latest git (apply the patch) and make run of BE tests we > have recently. > > Will keep you updated with reports then. > > > Thanks, > Garik. Done, All green with an exception of 1 failing tests - which is just a TODO on developers list (regarding promote product without syncing) - known stuff and reported on fedorahosted failed tests wiki. (nothing common with patch) So - patch seems "green" :) thanks a lot, Garik From lzap at redhat.com Mon Dec 19 13:55:19 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Mon, 19 Dec 2011 14:55:19 +0100 Subject: [katello-devel] Backed patch verification please In-Reply-To: <4EEF40EF.1010707@redhat.com> References: <4EEBA424.6060709@redhat.com> <4EEF1B70.4030806@redhat.com> <4EEF40EF.1010707@redhat.com> Message-ID: <20111219135519.GA2946@lzapx.brq.redhat.com> Ivan you mentioned the promoting of empty (unsynced) repo is maybe not correct and we should allow it. What is the status please? ps - If anybody apply the patch please report here ;-) lZ On Mon, Dec 19, 2011 at 02:49:35PM +0100, Garik Khachikyan wrote: > On 12/19/2011 12:09 PM, Garik Khachikyan wrote: > >On 12/16/2011 09:03 PM, Bryan Kearney wrote: > >>Can someone who has the backend test suite running try this > >>patch for me? Zeus tested it live and it works there, I want to > >>make sure there are no regressions. > >> > >>-- bk > >> > >> > >>_______________________________________________ > >>katello-devel mailing list > >>katello-devel at redhat.com > >>https://www.redhat.com/mailman/listinfo/katello-devel > >Hi Bryan, > > > >I'll take the latest git (apply the patch) and make run of BE > >tests we have recently. > > > >Will keep you updated with reports then. > > > > > >Thanks, > >Garik. > Done, > > All green with an exception of 1 failing tests - which is just a > TODO on developers list (regarding promote product without syncing) > - known stuff and reported on fedorahosted failed tests wiki. > (nothing common with patch) > > So - patch seems "green" :) > > thanks a lot, > Garik > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From inecas at redhat.com Mon Dec 19 14:20:55 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Mon, 19 Dec 2011 15:20:55 +0100 Subject: [katello-devel] Backed patch verification please In-Reply-To: <20111219135519.GA2946@lzapx.brq.redhat.com> References: <4EEBA424.6060709@redhat.com> <4EEF1B70.4030806@redhat.com> <4EEF40EF.1010707@redhat.com> <20111219135519.GA2946@lzapx.brq.redhat.com> Message-ID: <4EEF4847.7060104@redhat.com> On 12/19/2011 02:55 PM, Lukas Zapletal wrote: > Ivan you mentioned the promoting of empty (unsynced) repo is maybe not > correct and we should allow it. What is the status please? I mentioned it might be correct to promote unsynced repo. The status is here [1] and here [2]. The bug was on the Pulp side and with the latest CR it should be gone. Therefore I suggest to remove the conditions on repo to be synced before promoting. [1] - https://bugzilla.redhat.com/show_bug.cgi?id=765888 [2] - https://bugzilla.redhat.com/show_bug.cgi?id=760310 > > ps - If anybody apply the patch please report here ;-) > > lZ > > On Mon, Dec 19, 2011 at 02:49:35PM +0100, Garik Khachikyan wrote: >> On 12/19/2011 12:09 PM, Garik Khachikyan wrote: >>> On 12/16/2011 09:03 PM, Bryan Kearney wrote: >>>> Can someone who has the backend test suite running try this >>>> patch for me? Zeus tested it live and it works there, I want to >>>> make sure there are no regressions. >>>> >>>> -- bk >>>> >>>> >>>> _______________________________________________ >>>> katello-devel mailing list >>>> katello-devel at redhat.com >>>> https://www.redhat.com/mailman/listinfo/katello-devel >>> Hi Bryan, >>> >>> I'll take the latest git (apply the patch) and make run of BE >>> tests we have recently. >>> >>> Will keep you updated with reports then. >>> >>> >>> Thanks, >>> Garik. >> Done, >> >> All green with an exception of 1 failing tests - which is just a >> TODO on developers list (regarding promote product without syncing) >> - known stuff and reported on fedorahosted failed tests wiki. >> (nothing common with patch) >> >> So - patch seems "green" :) >> >> thanks a lot, >> Garik >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From lzap at redhat.com Mon Dec 19 15:14:02 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Mon, 19 Dec 2011 16:14:02 +0100 Subject: [katello-devel] Backed patch verification please In-Reply-To: <4EEF4847.7060104@redhat.com> References: <4EEBA424.6060709@redhat.com> <4EEF1B70.4030806@redhat.com> <4EEF40EF.1010707@redhat.com> <20111219135519.GA2946@lzapx.brq.redhat.com> <4EEF4847.7060104@redhat.com> Message-ID: <20111219151402.GB2946@lzapx.brq.redhat.com> Sounds like a plan. Will do. Thanks. On Mon, Dec 19, 2011 at 03:20:55PM +0100, Ivan Necas wrote: > On 12/19/2011 02:55 PM, Lukas Zapletal wrote: > >Ivan you mentioned the promoting of empty (unsynced) repo is maybe not > >correct and we should allow it. What is the status please? > I mentioned it might be correct to promote unsynced repo. The status > is here [1] and here [2]. The bug was on the Pulp side and with the > latest CR it should be gone. Therefore I suggest to remove the > conditions on repo to be synced before promoting. > > [1] - https://bugzilla.redhat.com/show_bug.cgi?id=765888 > [2] - https://bugzilla.redhat.com/show_bug.cgi?id=760310 > > > >ps - If anybody apply the patch please report here ;-) > > > >lZ > > > >On Mon, Dec 19, 2011 at 02:49:35PM +0100, Garik Khachikyan wrote: > >>On 12/19/2011 12:09 PM, Garik Khachikyan wrote: > >>>On 12/16/2011 09:03 PM, Bryan Kearney wrote: > >>>>Can someone who has the backend test suite running try this > >>>>patch for me? Zeus tested it live and it works there, I want to > >>>>make sure there are no regressions. > >>>> > >>>>-- bk > >>>> > >>>> > >>>>_______________________________________________ > >>>>katello-devel mailing list > >>>>katello-devel at redhat.com > >>>>https://www.redhat.com/mailman/listinfo/katello-devel > >>>Hi Bryan, > >>> > >>>I'll take the latest git (apply the patch) and make run of BE > >>>tests we have recently. > >>> > >>>Will keep you updated with reports then. > >>> > >>> > >>>Thanks, > >>>Garik. > >>Done, > >> > >>All green with an exception of 1 failing tests - which is just a > >>TODO on developers list (regarding promote product without syncing) > >>- known stuff and reported on fedorahosted failed tests wiki. > >>(nothing common with patch) > >> > >>So - patch seems "green" :) > >> > >>thanks a lot, > >>Garik > >> > >>_______________________________________________ > >>katello-devel mailing list > >>katello-devel at redhat.com > >>https://www.redhat.com/mailman/listinfo/katello-devel > > > -- > Regards, > > Ivan Necas > Red Hat Czech s.r.o. Brno > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From bkearney at redhat.com Mon Dec 19 15:20:51 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Mon, 19 Dec 2011 10:20:51 -0500 Subject: [katello-devel] Backed patch verification please In-Reply-To: <4EEF40EF.1010707@redhat.com> References: <4EEBA424.6060709@redhat.com> <4EEF1B70.4030806@redhat.com> <4EEF40EF.1010707@redhat.com> Message-ID: <4EEF5653.5010602@redhat.com> On 12/19/2011 08:49 AM, Garik Khachikyan wrote: > On 12/19/2011 12:09 PM, Garik Khachikyan wrote: >> On 12/16/2011 09:03 PM, Bryan Kearney wrote: >>> Can someone who has the backend test suite running try this patch for >>> me? Zeus tested it live and it works there, I want to make sure there >>> are no regressions. >>> >>> -- bk >>> >>> >>> _______________________________________________ >>> katello-devel mailing list >>> katello-devel at redhat.com >>> https://www.redhat.com/mailman/listinfo/katello-devel >> Hi Bryan, >> >> I'll take the latest git (apply the patch) and make run of BE tests we >> have recently. >> >> Will keep you updated with reports then. >> >> >> Thanks, >> Garik. > Done, > > All green with an exception of 1 failing tests - which is just a TODO on > developers list (regarding promote product without syncing) - known > stuff and reported on fedorahosted failed tests wiki. > (nothing common with patch) > > So - patch seems "green" :) > > thanks a lot, > Garik Thank you! - kb From mmccune at redhat.com Mon Dec 19 18:35:08 2011 From: mmccune at redhat.com (Mike McCune) Date: Mon, 19 Dec 2011 10:35:08 -0800 Subject: [katello-devel] tagging with tito Message-ID: <4EEF83DC.4010203@redhat.com> If you ever tag our packages please read: https://fedorahosted.org/katello/wiki/TitoTagging note that I mention there to never accept the auto-changelog and always 'git push && git push --tags' Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From lzap at redhat.com Tue Dec 20 08:42:21 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 20 Dec 2011 09:42:21 +0100 Subject: [katello-devel] Candlepin import error: unable to extract export archive Message-ID: <20111220084221.GB7278@lzapx.brq.redhat.com> Hello, few months ago I was hitting strange issue when Candlepin was not able to import any manifest. Today I hit it again and, what is great, I have a Beaker reproducer reserved. Handing over to the Candlepin team for investigation. Thanks. https://bugzilla.redhat.com/show_bug.cgi?id=769194 I'd appreciate quick look because the box is reserved for 99 hours from now on. Thank you. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Tue Dec 20 09:36:19 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 20 Dec 2011 10:36:19 +0100 Subject: [katello-devel] tagging with tito In-Reply-To: <4EEF83DC.4010203@redhat.com> References: <4EEF83DC.4010203@redhat.com> Message-ID: <20111220093619.GC7278@lzapx.brq.redhat.com> Is the only reason to prevent having changelog very long? Is it a real problem to keep changelog entries? We can always delete them afterwards by simple editing the SPEC file. This won't break tito, I tried it. I have a macro that deletes all e-mails. I usually delete all messages that are not much informative (Merged branch 'master'). I don't think its a big issue. Can I keep changelog items there? Btw I added one more rule on the wiki page: '''never rebase unpushed tag''': If you make a tag and locally rebase, it will break our repo. The rule is to push your work first and to do tito tag then in a shortest time window possible. LZ On Mon, Dec 19, 2011 at 10:35:08AM -0800, Michael McCune wrote: > If you ever tag our packages please read: > > https://fedorahosted.org/katello/wiki/TitoTagging > > note that I mention there to never accept the auto-changelog and > always 'git push && git push --tags' > > Mike > -- > Mike McCune > mmccune AT redhat.com > Red Hat Engineering | Portland, OR > Systems Management | 650.254.4248 > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Tue Dec 20 15:25:18 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 20 Dec 2011 16:25:18 +0100 Subject: [katello-devel] Yum - Metadata file does not match checksum Message-ID: <20111220152518.GD7278@lzapx.brq.redhat.com> Hello, if we trigger complete rebuild of katello we also rebuild and reupload all rubygems. We do this even when the content does not change, so if your yum already cache some of those files, it will likely fail next time. The workaround is simple if you see the "Metadata file does not match checksum" error: # yum clean all -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Tue Dec 20 15:26:47 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 20 Dec 2011 16:26:47 +0100 Subject: [katello-devel] Candlepin import error: unable to extract export archive In-Reply-To: <20111220084221.GB7278@lzapx.brq.redhat.com> References: <20111220084221.GB7278@lzapx.brq.redhat.com> Message-ID: <20111220152647.GE7278@lzapx.brq.redhat.com> As Bryan told me today, this is a bug in Candlepin which was fixed in 0.5.8. You can find it in brew. LZ On Tue, Dec 20, 2011 at 09:42:21AM +0100, Lukas Zapletal wrote: > Hello, > > few months ago I was hitting strange issue when Candlepin was not able > to import any manifest. > > Today I hit it again and, what is great, I have a Beaker reproducer > reserved. > > Handing over to the Candlepin team for investigation. Thanks. > > https://bugzilla.redhat.com/show_bug.cgi?id=769194 > > I'd appreciate quick look because the box is reserved for 99 hours > from now on. Thank you. > > -- > Later, > > Lukas Zapletal | E32E400A > RHN Satellite Engineering > Red Hat Czech s.r.o. Brno > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From bkearney at redhat.com Tue Dec 20 15:30:47 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Tue, 20 Dec 2011 10:30:47 -0500 Subject: [katello-devel] Candlepin import error: unable to extract export archive In-Reply-To: <20111220152647.GE7278@lzapx.brq.redhat.com> References: <20111220084221.GB7278@lzapx.brq.redhat.com> <20111220152647.GE7278@lzapx.brq.redhat.com> Message-ID: <4EF0AA27.2080503@redhat.com> On 12/20/2011 10:26 AM, Lukas Zapletal wrote: > As Bryan told me today, this is a bug in Candlepin which was fixed in > 0.5.8. You can find it in brew. Ad all public repos. -- bk From lzap at redhat.com Tue Dec 20 16:15:29 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Tue, 20 Dec 2011 17:15:29 +0100 Subject: [katello-devel] New activation keys registration process Message-ID: <20111220161529.GF7278@lzapx.brq.redhat.com> Hey, as I promised on the demo, here is the wikipage describing the new registration process: https://fedorahosted.org/katello/wiki/ActivationKeysDesign QAs: Please consult this page when verifying the new activation keys. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From mmccune at redhat.com Tue Dec 20 17:02:10 2011 From: mmccune at redhat.com (Mike McCune) Date: Tue, 20 Dec 2011 09:02:10 -0800 Subject: [katello-devel] tagging with tito In-Reply-To: <20111220093619.GC7278@lzapx.brq.redhat.com> References: <4EEF83DC.4010203@redhat.com> <20111220093619.GC7278@lzapx.brq.redhat.com> Message-ID: <4EF0BF92.5020800@redhat.com> On 12/20/2011 01:36 AM, Lukas Zapletal wrote: > Is the only reason to prevent having changelog very long? > > Is it a real problem to keep changelog entries? We can always delete > them afterwards by simple editing the SPEC file. This won't break tito, > I tried it. > > I have a macro that deletes all e-mails. I usually delete all messages > that are not much informative (Merged branch 'master'). I don't think > its a big issue. > > Can I keep changelog items there? > if you want to hand-edit the changelog entry with a good summary of the changes from git I'm totally fine. The more complex rule would be: "If you don't want to take the time to edit the automatic generated changelog entry to trim the merge commits and other useless noise just put in 'periodic rebuild', otherwise feel free to include the highlight commits for features and bugfixes" I'd rather keep the specfile cleaner and smaller than too verbose. > Btw I added one more rule on the wiki page: > > '''never rebase unpushed tag''': If you make a tag and locally rebase, > it will break our repo. The rule is to push your work first and to do > tito tag then in a shortest time window possible. +1, I forgot about that one. Mike From inecas at redhat.com Tue Dec 20 17:04:14 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Tue, 20 Dec 2011 18:04:14 +0100 Subject: [katello-devel] Beware of :id => false in our migrations unless HABTM Message-ID: <4EF0C00E.5060902@redhat.com> Hi all, Today I've hit and interesting bug in Rails that showed up in our project [1]. How to reproduce (katello-0.1.150-1): 1. create a table without primary id (:id => false in our migration) (system_tasks in our case) 2. create a model for that table (SystemTask) 3. use has_many to that model (in System has_many system_tasks) 4. create a record in SystemTask 5. try to destroy the assigned record (System) 6. enjoy NoMethodError: undefined method `eq' for nil:NilClass and really long stack trace. How to avoid: Besides previous hint not to use HABTM there is another rule of thumb: * Don't use :id => false* in migrations [1] - http://stackoverflow.com/questions/5085258/cant-destroy-record-in-many-to-many-relationship -- Ivan -------------- next part -------------- An HTML attachment was scrubbed... URL: From jrist at redhat.com Tue Dec 20 17:53:22 2011 From: jrist at redhat.com (Jason Rist) Date: Tue, 20 Dec 2011 10:53:22 -0700 Subject: [katello-devel] Beware of :id => false in our migrations unless HABTM In-Reply-To: <4EF0C00E.5060902@redhat.com> References: <4EF0C00E.5060902@redhat.com> Message-ID: <4EF0CB92.1070907@redhat.com> On 12/20/2011 10:04 AM, Ivan Ne?as wrote: > Hi all, > > Today I've hit and interesting bug in Rails that showed up in our > project [1]. > > How to reproduce (katello-0.1.150-1): > > 1. create a table without primary id (:id => false in our migration) > (system_tasks in our case) > 2. create a model for that table (SystemTask) > 3. use has_many to that model (in System has_many system_tasks) > 4. create a record in SystemTask > 5. try to destroy the assigned record (System) > 6. enjoy NoMethodError: undefined method `eq' for nil:NilClass and > really long stack trace. > > > How to avoid: > > Besides previous hint not to use HABTM there is another rule of thumb: > * > Don't use :id => false* in migrations > > [1] - > http://stackoverflow.com/questions/5085258/cant-destroy-record-in-many-to-many-relationship > > > -- Ivan > Thanks for the note, Ivan - very helpful! -J -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From ohadlevy at redhat.com Wed Dec 21 06:41:54 2011 From: ohadlevy at redhat.com (Ohad Levy) Date: Wed, 21 Dec 2011 01:41:54 -0500 (EST) Subject: [katello-devel] Beware of :id => false in our migrations unless HABTM In-Reply-To: <4EF0C00E.5060902@redhat.com> Message-ID: <3b63c26c-563d-4b35-857d-b8bb3b443145@zmail14.collab.prod.int.phx2.redhat.com> ----- Original Message ----- | Hi all, | Today I've hit and interesting bug in Rails that showed up in our | project [1]. Is that really a bug? if the many to many relationship is not HABTM but via a model, then you need a separate ID for it,.. i.e. when migrating from HABTM to has_may :though, you need to add an ID column. Ohad | How to reproduce (katello-0.1.150-1): | 1. create a table without primary id (:id => false in our migration) | (system_tasks in our case) | 2. create a model for that table (SystemTask) | 3. use has_many to that model (in System has_many system_tasks) | 4. create a record in SystemTask | 5. try to destroy the assigned record (System) | 6. enjoy NoMethodError: undefined method `eq' for nil:NilClass and | really long stack trace. | How to avoid: | Besides previous hint not to use HABTM there is another rule of | thumb: | Don't use :id => false in migrations | [1] - | http://stackoverflow.com/questions/5085258/cant-destroy-record-in-many-to-many-relationship | -- Ivan | _______________________________________________ | katello-devel mailing list | katello-devel at redhat.com | https://www.redhat.com/mailman/listinfo/katello-devel From inecas at redhat.com Wed Dec 21 09:22:50 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Wed, 21 Dec 2011 10:22:50 +0100 Subject: [katello-devel] Beware of :id => false in our migrations unless HABTM In-Reply-To: <3b63c26c-563d-4b35-857d-b8bb3b443145@zmail14.collab.prod.int.phx2.redhat.com> References: <3b63c26c-563d-4b35-857d-b8bb3b443145@zmail14.collab.prod.int.phx2.redhat.com> Message-ID: <4EF1A56A.1030705@redhat.com> On 12/21/2011 07:41 AM, Ohad Levy wrote: > Is that really a bug? if the many to many relationship is not HABTM > but via a model, then you need a separate ID for it,.. i.e. when > migrating from HABTM to has_may :though, you need to add an ID > column. > > Ohad The bug is at least, that it doesn't fail sooner, then at the destroy time, which is quite late for me. Even no warning is there. -- Ivan -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From jsherril at redhat.com Wed Dec 21 15:10:57 2011 From: jsherril at redhat.com (Justin Sherrill) Date: Wed, 21 Dec 2011 10:10:57 -0500 Subject: [katello-devel] Search merged to master Message-ID: <4EF1F701.4010604@redhat.com> Next time you pull, you'll want to: yum install elasticsearch rubygem-tire from your friendly community katello repo and start the elasticsearch service. Its recommended you do a wipe-all to allow katello to index the items. -Justin From jrist at redhat.com Wed Dec 21 15:22:55 2011 From: jrist at redhat.com (Jason Rist) Date: Wed, 21 Dec 2011 08:22:55 -0700 Subject: [katello-devel] Search merged to master In-Reply-To: <4EF1F701.4010604@redhat.com> References: <4EF1F701.4010604@redhat.com> Message-ID: <4EF1F9CF.1010708@redhat.com> On 12/21/2011 08:10 AM, Justin Sherrill wrote: > Next time you pull, you'll want to: > > > yum install elasticsearch rubygem-tire > > from your friendly community katello repo and start the elasticsearch > service. > > Its recommended you do a wipe-all to allow katello to index the items. > > -Justin > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel AWESOME -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From gkhachik at redhat.com Wed Dec 21 15:27:30 2011 From: gkhachik at redhat.com (Garik Khachikyan) Date: Wed, 21 Dec 2011 16:27:30 +0100 Subject: [katello-devel] Search merged to master In-Reply-To: <4EF1F701.4010604@redhat.com> References: <4EF1F701.4010604@redhat.com> Message-ID: <4EF1FAE2.804@redhat.com> On 12/21/2011 04:10 PM, Justin Sherrill wrote: > Next time you pull, you'll want to: > > > yum install elasticsearch rubygem-tire > > from your friendly community katello repo and start the elasticsearch > service. > > Its recommended you do a wipe-all to allow katello to index the items. > > -Justin > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel Hey Justin, So is it a thing we need (or have to) add in Katello installer I guess? thanks, Garik. From jsherril at redhat.com Wed Dec 21 15:29:06 2011 From: jsherril at redhat.com (Justin Sherrill) Date: Wed, 21 Dec 2011 10:29:06 -0500 Subject: [katello-devel] Search merged to master In-Reply-To: <4EF1FAE2.804@redhat.com> References: <4EF1F701.4010604@redhat.com> <4EF1FAE2.804@redhat.com> Message-ID: <4EF1FB42.50205@redhat.com> On 12/21/2011 10:27 AM, Garik Khachikyan wrote: > On 12/21/2011 04:10 PM, Justin Sherrill wrote: >> Next time you pull, you'll want to: >> >> >> yum install elasticsearch rubygem-tire >> >> from your friendly community katello repo and start the elasticsearch >> service. >> >> Its recommended you do a wipe-all to allow katello to index the items. >> >> -Justin >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel > Hey Justin, > > So is it a thing we need (or have to) add in Katello installer I guess? > > thanks, > Garik. The installer should already take care of it. As soon as regenerate the community repos, any install off of those should pick it up just fine. -Justin From mmccune at redhat.com Wed Dec 21 16:42:02 2011 From: mmccune at redhat.com (Mike McCune) Date: Wed, 21 Dec 2011 08:42:02 -0800 Subject: [katello-devel] Search merged to master In-Reply-To: <4EF1FB42.50205@redhat.com> References: <4EF1F701.4010604@redhat.com> <4EF1FAE2.804@redhat.com> <4EF1FB42.50205@redhat.com> Message-ID: <4EF20C5A.9020700@redhat.com> On 12/21/2011 07:29 AM, Justin Sherrill wrote: > On 12/21/2011 10:27 AM, Garik Khachikyan wrote: >> On 12/21/2011 04:10 PM, Justin Sherrill wrote: >>> Next time you pull, you'll want to: >>> >>> >>> yum install elasticsearch rubygem-tire >>> >>> from your friendly community katello repo and start the elasticsearch >>> service. >>> >>> Its recommended you do a wipe-all to allow katello to index the items. >>> >>> -Justin >>> >>> _______________________________________________ >>> katello-devel mailing list >>> katello-devel at redhat.com >>> https://www.redhat.com/mailman/listinfo/katello-devel >> Hey Justin, >> >> So is it a thing we need (or have to) add in Katello installer I guess? >> >> thanks, >> Garik. > The installer should already take care of it. As soon as regenerate the > community repos, any install off of those should pick it up just fine. > > -Justin > that is correct, there is a new puppet module that ensures elasticsearch is installed and running. Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From inecas at redhat.com Wed Dec 21 17:57:17 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Wed, 21 Dec 2011 18:57:17 +0100 Subject: [katello-devel] Search merged to master In-Reply-To: <4EF1F9CF.1010708@redhat.com> References: <4EF1F701.4010604@redhat.com> <4EF1F9CF.1010708@redhat.com> Message-ID: <4EF21DFD.2060704@redhat.com> On 12/21/2011 04:22 PM, Jason Rist wrote: > On 12/21/2011 08:10 AM, Justin Sherrill wrote: >> Next time you pull, you'll want to: >> >> >> yum install elasticsearch rubygem-tire >> >> from your friendly community katello repo and start the elasticsearch >> service. >> >> Its recommended you do a wipe-all to allow katello to index the items. Fedora 15 repo seems broken: http://repos.fedorapeople.org/repos/katello/katello/fedora-15/x86_64/ only one package is there: elasticsearch >> >> -Justin >> >> _______________________________________________ >> katello-devel mailing list >> katello-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/katello-devel > AWESOME > -- Regards, Ivan Necas Red Hat Czech s.r.o. Brno From mmccune at redhat.com Wed Dec 21 18:10:43 2011 From: mmccune at redhat.com (Mike McCune) Date: Wed, 21 Dec 2011 10:10:43 -0800 Subject: [katello-devel] Search merged to master In-Reply-To: <4EF21DFD.2060704@redhat.com> References: <4EF1F701.4010604@redhat.com> <4EF1F9CF.1010708@redhat.com> <4EF21DFD.2060704@redhat.com> Message-ID: <4EF22123.7050808@redhat.com> On 12/21/2011 09:57 AM, Ivan Ne?as wrote: > On 12/21/2011 04:22 PM, Jason Rist wrote: >> On 12/21/2011 08:10 AM, Justin Sherrill wrote: >>> Next time you pull, you'll want to: >>> >>> >>> yum install elasticsearch rubygem-tire >>> >>> from your friendly community katello repo and start the elasticsearch >>> service. >>> >>> Its recommended you do a wipe-all to allow katello to index the items. > Fedora 15 repo seems broken: > http://repos.fedorapeople.org/repos/katello/katello/fedora-15/x86_64/ > > only one package is there: elasticsearch should be fixed by the time most people read this -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From inecas at redhat.com Wed Dec 21 18:33:32 2011 From: inecas at redhat.com (=?UTF-8?B?SXZhbiBOZcSNYXM=?=) Date: Wed, 21 Dec 2011 19:33:32 +0100 Subject: [katello-devel] katello-agent cli support in master Message-ID: <4EF2267C.1060909@redhat.com> Hi all, I've merged the work on katello-agent cli support. After the next tag it will be available in RPM. New commands in cli system packages --install system packages --update system packages --update --all system packages --remove system packages --install_group system packages --remove_group More values you can separate with comma. The API looks like this: # install packages POST /systems/:system_uuid/packages {:packages => ['zsh','bash']} # install package groups POST /systems/:system_uuid/packages {:groups => ['Editors']} # remove packages DELETE /systems/:system_uuid/packages {:packages => ['zsh','bash']} # remove package groups DELETE /systems/:system_uuid/packages {:groups => ['Editors']} # update packages PUT /systems/:system_uuid/packages {:packages => ['zsh','bash']} All calls return info about the created task. This tasks are asynchronous. You can use CTRL+C to close the terminal and use task id to check the status with: system task --id TASK-UUID-FROM-THE-COMMAND (API: . GET /systems/tasks/:task_uuid ) you can also show all tasks for org/environment/system using system tasks --org [--environment] [--name] (API: . GET /organizations/:org_id/systems/tasks ) I've also create bunch of cli system tests, that test the whole functionality of katello-agent. It runs if the katello-agent is installed on the machine. It checks whether packages and package groups get installed, removed and updated. Garik: For ci machine running system tests, see [1] for proper setting of the client. [1] - https://fedorahosted.org/katello/wiki/KatelloAgent#SSLHowTo -- Ivan From jsherril at redhat.com Wed Dec 21 20:49:35 2011 From: jsherril at redhat.com (Justin Sherrill) Date: Wed, 21 Dec 2011 15:49:35 -0500 Subject: [katello-devel] Quick note on elastic search Message-ID: <4EF2465F.9000800@redhat.com> Apparently elastic search enjoys clustering itself by default. We're disabling that in the installer, but in the meantime I HIGHLY recommend you add : discovery.zen.ping.multicast.enabled: false or else your elastic search server will start communicating with other es servers on your subnet :) -Justin From mmccune at redhat.com Wed Dec 21 20:53:35 2011 From: mmccune at redhat.com (Mike McCune) Date: Wed, 21 Dec 2011 12:53:35 -0800 Subject: [katello-devel] Quick note on elastic search In-Reply-To: <4EF2465F.9000800@redhat.com> References: <4EF2465F.9000800@redhat.com> Message-ID: <4EF2474F.1030503@redhat.com> On 12/21/2011 12:49 PM, Justin Sherrill wrote: > Apparently elastic search enjoys clustering itself by default. We're > disabling that in the installer, but in the meantime I HIGHLY recommend > you add : > > discovery.zen.ping.multicast.enabled: false > > > or else your elastic search server will start communicating with other > es servers on your subnet :) > > -Justin > to be clear, append that value to the end of your: /etc/elasticsearch/elasticsearch.yml Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From mmccune at redhat.com Wed Dec 21 22:20:20 2011 From: mmccune at redhat.com (Mike McCune) Date: Wed, 21 Dec 2011 14:20:20 -0800 Subject: [katello-devel] Quick note on elastic search In-Reply-To: <4EF2474F.1030503@redhat.com> References: <4EF2465F.9000800@redhat.com> <4EF2474F.1030503@redhat.com> Message-ID: <4EF25BA4.7060104@redhat.com> On 12/21/2011 12:53 PM, Mike McCune wrote: > On 12/21/2011 12:49 PM, Justin Sherrill wrote: >> Apparently elastic search enjoys clustering itself by default. We're >> disabling that in the installer, but in the meantime I HIGHLY recommend >> you add : >> >> discovery.zen.ping.multicast.enabled: false >> >> >> or else your elastic search server will start communicating with other >> es servers on your subnet :) >> >> -Justin >> > > to be clear, append that value to the end of your: > > /etc/elasticsearch/elasticsearch.yml > > Mike latest version of katello-configure has this disabled by default now. Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From ohadlevy at redhat.com Thu Dec 22 07:48:31 2011 From: ohadlevy at redhat.com (Ohad Levy) Date: Thu, 22 Dec 2011 02:48:31 -0500 (EST) Subject: [katello-devel] Non-ar search discussion In-Reply-To: <4EE8A113.1000909@redhat.com> Message-ID: <2390b4dd-60d4-423d-88ad-dbdbaa99a0e5@zmail14.collab.prod.int.phx2.redhat.com> Did anyone paid for the pro episodes? do you think its worth it? Ohad ----- Original Message ----- | Btw. regarding the %subject% there are recently 3 episodes on | RailsCasts | that might be helpful. | | http://railscasts.com/episodes/306-elasticsearch-part-1 | http://railscasts.com/episodes/307-elasticsearch-part-2 | http://railscasts.com/episodes/120-thinking-sphinx-revised | | -- Ivan | | On 12/09/2011 03:35 PM, Ivan Ne?as wrote: | > On 12/09/2011 03:23 PM, Justin Sherrill wrote: | >> Have you looked at elastic search? Its the one I'm currently | >> favouring and the one that delta cloud was seriously considering | >> (they ended up pushing full text search for v2). It is very | >> lightweight and simple and from a quick glance all its | >> dependencies | >> are already in fedora (although not all at the correct version). | >> | >> -Justin | > | > Just a note, a Ruby client [1] for the ES is maintained by a Czech | > developer. For Lukas - ES is based on Lucene, with nice RESTful | > Interface. | > | > [1] - https://github.com/karmi/tire | > | > _______________________________________________ | > katello-devel mailing list | > katello-devel at redhat.com | > https://www.redhat.com/mailman/listinfo/katello-devel | | | -- | Regards, | | Ivan Necas | Red Hat Czech s.r.o. Brno | | _______________________________________________ | katello-devel mailing list | katello-devel at redhat.com | https://www.redhat.com/mailman/listinfo/katello-devel | From lzap at redhat.com Thu Dec 22 16:17:06 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Thu, 22 Dec 2011 17:17:06 +0100 Subject: [katello-devel] Katello certificates Message-ID: <20111222161706.GD3939@lzapx.brq.redhat.com> Hello, I disabled the temporary fix for NSS certificate store generation in Puppet. It seems to be fixed now on RHEL 6.2 (QA verified). Current version of Katello installer deploys several certificates, which are currently not uses by this backend engine. The complete list can be found here: https://fedorahosted.org/katello/wiki/Certificates#Katellopuppetinstallergenerates (chapter "Katello puppet installer generates") The idea is Katello (and it's installer) will generate all necessary certificates for all backend engines. All will be signed with the only one CA (also generated by the Katello installer). I have created two tracking RHBZ for implementing this big configuration change as I expect changes on the backend engines: Candlepin - https://bugzilla.redhat.com/show_bug.cgi?id=769887 Pulp - https://bugzilla.redhat.com/show_bug.cgi?id=769888 I would like to ask folks from both teams to help me with setting new certificates (or finding other certificates that needs to be deployed), testing the setup and also with orchestrating changes on their side if necessary. I guess we will start working on this next year, but I'd be happy for opinions or recommendations. -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From mmccune at redhat.com Thu Dec 22 20:22:46 2011 From: mmccune at redhat.com (Mike McCune) Date: Thu, 22 Dec 2011 12:22:46 -0800 Subject: [katello-devel] getting rid of Gemfile.lock in git Message-ID: <4EF39196.2060001@redhat.com> I don't see a compelling reason to keep this file in git anymore, in production with RPMs we delete the Gemfile.lock and regenerate based on the RPMs installed on the system. Can we just nuke it from git and have devs run 'bundle install --local' to regenerate the proper Gemfile.lock? Mike -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From jrist at redhat.com Thu Dec 22 20:23:34 2011 From: jrist at redhat.com (Jason Rist) Date: Thu, 22 Dec 2011 13:23:34 -0700 Subject: [katello-devel] getting rid of Gemfile.lock in git In-Reply-To: <4EF39196.2060001@redhat.com> References: <4EF39196.2060001@redhat.com> Message-ID: <4EF391C6.6010100@redhat.com> On Thu 22 Dec 2011 01:22:46 PM MST, Mike McCune wrote: > I don't see a compelling reason to keep this file in git anymore, in > production with RPMs we delete the Gemfile.lock and regenerate based > on the RPMs installed on the system. > > Can we just nuke it from git and have devs run 'bundle install > --local' to regenerate the proper Gemfile.lock? > > Mike +1 -- Jason E. Rist Senior Software Engineer Systems Management and Cloud Enablement Red Hat, Inc. +1.919.754.4048 Freenode: jrist From bkearney at redhat.com Thu Dec 22 20:53:59 2011 From: bkearney at redhat.com (Bryan Kearney) Date: Thu, 22 Dec 2011 15:53:59 -0500 Subject: [katello-devel] getting rid of Gemfile.lock in git In-Reply-To: <4EF391C6.6010100@redhat.com> References: <4EF39196.2060001@redhat.com> <4EF391C6.6010100@redhat.com> Message-ID: <4EF398E7.3030908@redhat.com> On 12/22/2011 03:23 PM, Jason Rist wrote: > On Thu 22 Dec 2011 01:22:46 PM MST, Mike McCune wrote: >> I don't see a compelling reason to keep this file in git anymore, in >> production with RPMs we delete the Gemfile.lock and regenerate based >> on the RPMs installed on the system. >> >> Can we just nuke it from git and have devs run 'bundle install >> --local' to regenerate the proper Gemfile.lock? >> >> Mike > > +1 > I dont know the answer.. but does it give us an indication of new dependenceis being added? -- bk From mmccune at redhat.com Thu Dec 22 21:44:32 2011 From: mmccune at redhat.com (Mike McCune) Date: Thu, 22 Dec 2011 13:44:32 -0800 Subject: [katello-devel] getting rid of Gemfile.lock in git In-Reply-To: <4EF398E7.3030908@redhat.com> References: <4EF39196.2060001@redhat.com> <4EF391C6.6010100@redhat.com> <4EF398E7.3030908@redhat.com> Message-ID: <4EF3A4C0.3050000@redhat.com> On 12/22/2011 12:53 PM, Bryan Kearney wrote: > On 12/22/2011 03:23 PM, Jason Rist wrote: >> On Thu 22 Dec 2011 01:22:46 PM MST, Mike McCune wrote: >>> I don't see a compelling reason to keep this file in git anymore, in >>> production with RPMs we delete the Gemfile.lock and regenerate based >>> on the RPMs installed on the system. >>> >>> Can we just nuke it from git and have devs run 'bundle install >>> --local' to regenerate the proper Gemfile.lock? >>> >>> Mike >> >> +1 >> > I dont know the answer.. but does it give us an indication of new > dependenceis being added? > > -- bk > no, we manage that with brew and our internal gem/rpm repository process, the Gemfile.lock is the *last* thing to know about deps. -- Mike McCune mmccune AT redhat.com Red Hat Engineering | Portland, OR Systems Management | 650.254.4248 From lzap at redhat.com Fri Dec 23 07:41:10 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 23 Dec 2011 08:41:10 +0100 Subject: [katello-devel] getting rid of Gemfile.lock in git In-Reply-To: <4EF39196.2060001@redhat.com> References: <4EF39196.2060001@redhat.com> Message-ID: <20111223074110.GA2822@lzapx.brq.redhat.com> Absolutely! Plus to add it to the .gitignore... LZ On Thu, Dec 22, 2011 at 12:22:46PM -0800, Michael McCune wrote: > I don't see a compelling reason to keep this file in git anymore, in > production with RPMs we delete the Gemfile.lock and regenerate based > on the RPMs installed on the system. > > Can we just nuke it from git and have devs run 'bundle install > --local' to regenerate the proper Gemfile.lock? > > Mike > -- > Mike McCune > mmccune AT redhat.com > Red Hat Engineering | Portland, OR > Systems Management | 650.254.4248 > > _______________________________________________ > katello-devel mailing list > katello-devel at redhat.com > https://www.redhat.com/mailman/listinfo/katello-devel -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno From lzap at redhat.com Fri Dec 23 10:32:55 2011 From: lzap at redhat.com (Lukas Zapletal) Date: Fri, 23 Dec 2011 11:32:55 +0100 Subject: [katello-devel] Permissions unit test coverage Message-ID: <20111223103255.GC2822@lzapx.brq.redhat.com> Hello, I created a simple code that allows us to track unit test coverage of our permissions: https://fedorahosted.org/katello/wiki/PermissionUnitCoverage This report show all controllers and actions which are tested with it_should_behave_like "protected action" shared example. Actions are tested twice - one time with some permissions assigned (see the 3rd column), second time without any (or particular) permissions (this table does not cover that as we deny access by default). How to create this report: PERMISSION_COVERAGE=perm_coverage.log rake spec Basically it's for me, but feel free to regenerate if you like it ;-) Merry Xmas! LZ -- Later, Lukas Zapletal | E32E400A RHN Satellite Engineering Red Hat Czech s.r.o. Brno