[Pulp-list] Runaway Pulp Tasks

[Dennis Kliban]

I think you need to use rabbitmqctl purge_queue[0] command to purge all the
messages for the resource manager queue. I believe you want to do this
while the resource_manager is offline and the pulp_workers are offline.

http://www.rabbitmq.com/man/rabbitmqctl.1.man.html

On Mon, Nov 13, 2017 at 11:11 AM, Dustin McNabb <dustin.mcnabb at gmail.com>
wrote:

> That only works for completed tasks. Mine are all in Waiting state.
>
> Thanks
> Dustin
>
> On Nov 13, 2017, at 8:03 AM, Deej Howard <Deej.Howard at neulion.com> wrote:
>
>                 I assume you’ve already tried using “pulp-admin tasks
> purge”, right (ref: https://www.mankier.com/1/pulp-admin#Tasks-Purging)?
>
>
>
> *From:* pulp-list-bounces at redhat.com [mailto:pulp-list-bounces at redhat.com]
> *On Behalf Of *Dustin McNabb
> *Sent:* Monday, November 13, 2017 9:00 AM
> *To:* Dennis Kliban <dkliban at redhat.com>
> *Cc:* pulp-list <pulp-list at redhat.com>
> *Subject:* Re: [Pulp-list] Runaway Pulp Tasks
>
>
>
> We are using rabbitMQ rather than qpidd, but I’ve stopped and started all
> of the relevant services to no avail.
>
> Thanks
>
> Dustin
>
>
> On Nov 11, 2017, at 7:45 AM, Dennis Kliban <dkliban at redhat.com> wrote:
>
> You can stop pulp_resource_manager, pulp_workers, and qpidd. Then start
> qpidd, pulp_resource_manager, and pulp_workers. The workers will mark the
> tasks as canceled in the database when they start and qpid should drain the
> queue when it is restarted.
>
>
>
> On Fri, Nov 10, 2017 at 6:33 PM, Dustin McNabb <dustin.mcnabb at gmail.com>
> wrote:
>
> Anyone ever encountered anything like this? I found 67K pulp tasks pending
> on one of my pulp servers today. I’m running pulp 2.10.3-1 (I know I know
> I’m fixing that) with mongoDB 2.6 on RHEL7 with a couple hundred yum repos
> and one python repo. We recently developed a config mgmt. state to manage
> the pulp repos on all our pulp servers, and that appears to have caused
> this issue by submitting a large number of pulp tasks in a short period and
> then trying again 30 minutes later. My question now, is how can I kill so
> many pulp tasks in a more efficient manner than the for loop I’m using?
>
>
>
> # A whole lotta pulp tasks
>
> [root at pulp-server :~]# pulp-admin tasks list |grep 'Task Id' |wc -l
>
> 67016
>
>
>
> # A for loop to generate a list of the task IDs and cancel one at a time.
>
> for n in `pulp-admin tasks list |grep 'Task Id' |awk '{print $NF}'`; do
> echo $n; pulp-admin tasks cancel --task-id $n; done
>
>
>
> Thanks
>
> Dustin
>
>
> _______________________________________________
> 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: <http://listman.redhat.com/archives/pulp-list/attachments/20171113/01473b1a/attachment.htm>