[Pulp-list] Pulp-list Digest, Vol 61, Issue 8

Gregory Komissarov gregory.komissarov at gmail.com
Sun Dec 7 08:15:20 UTC 2014 

I found the workaround that's ugly but worjing:
in repo
folder /var/lib/pulp/published/yum/master/yum_distributor/repo_qa_rhel6
remove the folder with metadata and symlinls rm -Rf 1417939672.39
and call publish after pulp-admin rpm repo publish run
--repo-id=repo_qa_rhel6

after this steps all works fine, all prev published pkgs seen to all yum
clients.
currentlly I'm trying to find in source code write place to paste folder
remove action to do such trick automatically.

hardcode only_publish_directory_contents=False in publish_step.py didn't
help.

On Sat, Dec 6, 2014 at 8:00 PM, <pulp-list-request at redhat.com> wrote:

> Send Pulp-list mailing list submissions to
>         pulp-list at redhat.com
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://www.redhat.com/mailman/listinfo/pulp-list
> or, via email, send a message with subject or body 'help' to
>         pulp-list-request at redhat.com
>
> You can reach the person managing the list at
>         pulp-list-owner at redhat.com
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Pulp-list digest..."
>
>
> Today's Topics:
>
>    1. quick functionality question about repo metadata (Jon Shanks)
>    2. Re: quick functionality question about repo metadata
>       (Barnaby Court)
>    3. Re: quick functionality question about repo metadata (Jon Shanks)
>    4. Re: quick functionality question about repo metadata
>       (Barnaby Court)
>    5. Re: Consumer repo bind validation (Jeff Ortel)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 5 Dec 2014 17:16:53 +0000
> From: Jon Shanks <jon.shanks at gmail.com>
> To: "Pulp-list at redhat.com" <pulp-list at redhat.com>
> Subject: [Pulp-list] quick functionality question about repo metadata
> Message-ID:
>         <
> CA+HXFk8UUtcXLHTX9QXbheseVn-LTBagy8U90u6wZyeUANGyQA at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Hi,
>
> I can't tell if there is a bug or if it's a new feature in pulp between 2.4
> and 2.5 but the publish repo seems to create a new timestamped directory
> for the publish however the repodata directory still seems to retain other
> checksums and grows with each run i.e. the primary, filelist, updateinfo
> and other xml metadata seems to retain the previous meta information from
> the previous runs.
>
> Is there some reason for this?
>
> Thanks
>
> Jon
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <
> https://www.redhat.com/archives/pulp-list/attachments/20141205/a2fa7ddb/attachment.html
> >
>
> ------------------------------
>
> Message: 2
> Date: Fri, 5 Dec 2014 12:25:43 -0500 (EST)
> From: Barnaby Court <bcourt at redhat.com>
> To: Jon Shanks <jon.shanks at gmail.com>
> Cc: "Pulp-list at redhat.com" <pulp-list at redhat.com>
> Subject: Re: [Pulp-list] quick functionality question about repo
>         metadata
> Message-ID:
>         <1744369251.9537576.1417800343254.JavaMail.zimbra at redhat.com>
> Content-Type: text/plain; charset=utf-8
>
> Hi, This is by design.  If files have not been removed from the repo we
> perform an incremental publish which does not remove the old files.  If an
> RPM is removed from the repo we will perform a clean publish.  The
> repomd.xml will only reference the new files so yum should play nice.
>
> -Barnaby
>
> ----- Original Message -----
> From: "Jon Shanks" <jon.shanks at gmail.com>
> To: "Pulp-list at redhat.com" <pulp-list at redhat.com>
> Sent: Friday, December 5, 2014 12:16:53 PM
> Subject: [Pulp-list] quick functionality question about repo metadata
>
> Hi,
>
> I can't tell if there is a bug or if it's a new feature in pulp between
> 2.4 and 2.5 but the publish repo seems to create a new timestamped
> directory for the publish however the repodata directory still seems to
> retain other checksums and grows with each run i.e. the primary, filelist,
> updateinfo and other xml metadata seems to retain the previous meta
> information from the previous runs.
>
> Is there some reason for this?
>
> Thanks
>
> Jon
>
> _______________________________________________
> Pulp-list mailing list
> Pulp-list at redhat.com
> https://www.redhat.com/mailman/listinfo/pulp-list
>
>
>
> ------------------------------
>
> Message: 3
> Date: Fri, 5 Dec 2014 17:30:14 +0000
> From: Jon Shanks <jon.shanks at gmail.com>
> To: Barnaby Court <bcourt at redhat.com>
> Cc: "Pulp-list at redhat.com" <pulp-list at redhat.com>
> Subject: Re: [Pulp-list] quick functionality question about repo
>         metadata
> Message-ID:
>         <
> CA+HXFk_wbvqYtdui+SyKTpsV3nZzp6e-w8cjDaG25M3JAPPycA at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Yeh we are having problems where new RPM's are not always showing and it
> isn't to do with the meta cache local on the box as we do a yum clean all
> each time, sometimes they show and sometimes it's still showing only the
> old version when we push a new file up and publish. It seems intermittent
> and has only started happening since we upgraded 2.5 it has been fine for
> months on 2.4 and i noticed that this looks like a new feature
> implementation.
>
> i can investigate further and raise a bug.
>
> On 5 December 2014 at 17:25, Barnaby Court <bcourt at redhat.com> wrote:
>
> > Hi, This is by design.  If files have not been removed from the repo we
> > perform an incremental publish which does not remove the old files.  If
> an
> > RPM is removed from the repo we will perform a clean publish.  The
> > repomd.xml will only reference the new files so yum should play nice.
> >
> > -Barnaby
> >
> > ----- Original Message -----
> > From: "Jon Shanks" <jon.shanks at gmail.com>
> > To: "Pulp-list at redhat.com" <pulp-list at redhat.com>
> > Sent: Friday, December 5, 2014 12:16:53 PM
> > Subject: [Pulp-list] quick functionality question about repo metadata
> >
> > Hi,
> >
> > I can't tell if there is a bug or if it's a new feature in pulp between
> > 2.4 and 2.5 but the publish repo seems to create a new timestamped
> > directory for the publish however the repodata directory still seems to
> > retain other checksums and grows with each run i.e. the primary,
> filelist,
> > updateinfo and other xml metadata seems to retain the previous meta
> > information from the previous runs.
> >
> > Is there some reason for this?
> >
> > Thanks
> >
> > Jon
> >
> > _______________________________________________
> > Pulp-list mailing list
> > Pulp-list at redhat.com
> > https://www.redhat.com/mailman/listinfo/pulp-list
> >
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <
> https://www.redhat.com/archives/pulp-list/attachments/20141205/6756c65c/attachment.html
> >
>
> ------------------------------
>
> Message: 4
> Date: Fri, 5 Dec 2014 12:55:08 -0500 (EST)
> From: Barnaby Court <bcourt at redhat.com>
> To: Jon Shanks <jon.shanks at gmail.com>
> Cc: "Pulp-list at redhat.com" <pulp-list at redhat.com>
> Subject: Re: [Pulp-list] quick functionality question about repo
>         metadata
> Message-ID:
>         <1868744373.9553390.1417802108073.JavaMail.zimbra at redhat.com>
> Content-Type: text/plain; charset=utf-8
>
> Jon, you are not the only person that has been seeing this behavior.  I
> have not been able to duplicate locally so far but am continuing to
> investigate.  If you track down the exact nature of the issue by all means
> please let me know and I'll get a fix in ASAP.  If you clean out the
> previously published files, does yum work properly in that case?  They
> should be ignored since they are not referenced in the repomd.xml but I'm
> wondering if yum is doing something strange.
>
> -Barnaby
>
> ----- Original Message -----
> From: "Jon Shanks" <jon.shanks at gmail.com>
> To: "Barnaby Court" <bcourt at redhat.com>
> Cc: "Pulp-list at redhat.com" <pulp-list at redhat.com>
> Sent: Friday, December 5, 2014 12:30:14 PM
> Subject: Re: [Pulp-list] quick functionality question about repo metadata
>
> Yeh we are having problems where new RPM's are not always showing and it
> isn't to do with the meta cache local on the box as we do a yum clean all
> each time, sometimes they show and sometimes it's still showing only the
> old version when we push a new file up and publish. It seems intermittent
> and has only started happening since we upgraded 2.5 it has been fine for
> months on 2.4 and i noticed that this looks like a new feature
> implementation.
>
> i can investigate further and raise a bug.
>
> On 5 December 2014 at 17:25, Barnaby Court <bcourt at redhat.com> wrote:
>
> > Hi, This is by design.  If files have not been removed from the repo we
> > perform an incremental publish which does not remove the old files.  If
> an
> > RPM is removed from the repo we will perform a clean publish.  The
> > repomd.xml will only reference the new files so yum should play nice.
> >
> > -Barnaby
> >
> > ----- Original Message -----
> > From: "Jon Shanks" <jon.shanks at gmail.com>
> > To: "Pulp-list at redhat.com" <pulp-list at redhat.com>
> > Sent: Friday, December 5, 2014 12:16:53 PM
> > Subject: [Pulp-list] quick functionality question about repo metadata
> >
> > Hi,
> >
> > I can't tell if there is a bug or if it's a new feature in pulp between
> > 2.4 and 2.5 but the publish repo seems to create a new timestamped
> > directory for the publish however the repodata directory still seems to
> > retain other checksums and grows with each run i.e. the primary,
> filelist,
> > updateinfo and other xml metadata seems to retain the previous meta
> > information from the previous runs.
> >
> > Is there some reason for this?
> >
> > Thanks
> >
> > Jon
> >
> > _______________________________________________
> > Pulp-list mailing list
> > Pulp-list at redhat.com
> > https://www.redhat.com/mailman/listinfo/pulp-list
> >
>
>
>
> ------------------------------
>
> Message: 5
> Date: Fri, 05 Dec 2014 16:36:59 -0600
> From: Jeff Ortel <jortel at redhat.com>
> To: pulp-list at redhat.com
> Subject: Re: [Pulp-list] Consumer repo bind validation
> Message-ID: <5482338B.4090309 at redhat.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
> Hey Jon,
>
> The REST API [1] for binding a consumer to a repository returns a task
> that can be used to track the implementation of the bind request by the
> agent on the consumer.  When that task has completed, the bind work on
> the consumer has been completed.  The state of that task is a direct
> reflection of the state of that bind request to the agent.
>
> Hope this helps.
>
> -jeff
>
> [1]
>
> http://pulp-dev-guide.readthedocs.org/en/latest/integration/rest-api/consumer/bind.html
>
> On 12/05/2014 08:16 AM, Jon Shanks wrote:
> > Hi,
> >
> > I was wondering if there is some way to determine whether or not the
> > task is completed from a consumer end when it binds to a repo. In terms
> > of automation, when i bind a consumer, in this case via puppet, i have
> > had to put a sleep in as there is no real way for me to determine when
> > the task is completed from the server side, to then know to proceed with
> > other elements of the configuration.
> >
> > Without any real feature in there for awareness of task completion or
> > with it still running asynchronously it's hard to coordinate tasks
> > around the succession of repository creation on a node.
> >
> > If possible, i'd like to raise a feature if this doesn't exist, but not
> > sure where bugs / features get raised?
> >
> > Thanks
> >
> > Jon
> >
> >
> > _______________________________________________
> > Pulp-list mailing list
> > Pulp-list at redhat.com
> > https://www.redhat.com/mailman/listinfo/pulp-list
> >
>
>
>
> ------------------------------
>
> _______________________________________________
> Pulp-list mailing list
> Pulp-list at redhat.com
> https://www.redhat.com/mailman/listinfo/pulp-list
>
> End of Pulp-list Digest, Vol 61, Issue 8
> ****************************************
>



-- 
With great regards
Gregory Komissarov
Mob +7 905 288 4181
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/pulp-list/attachments/20141207/ac5efe70/attachment.htm>

More information about the Pulp-list mailing list