Thanks Mike. This works. Set the timeout to 120s. The query of all packages in rhel7 actually took less than a minute. Default 30s seems a little too short.
From: [email protected] At: 02/04/20 13:32:26To: [email protected] Cc: Bin Li (BLOOMBERG/ 120 PARK ) , [email protected] Subject: Re: [Pulp-list] pulp 3 list packages Hi Bin, Did this work for you? if so, it sounds like we should look into always having ansible-pulp / the unit file pass a value to gunicorn (and make it configurable.) -Mike On Fri, Jan 31, 2020 at 10:27 AM Dennis Kliban <[email protected]> wrote: This can be modified using the systemd unit file[0]. Gunicorn takes a -t argument. It defaults to 30 seconds. [0] https://github.com/pulp/ansible-pulp/blob/master/roles/pulp/templates/pulpcore-api.service.j2#L13 On Fri, Jan 31, 2020 at 9:26 AM Bin Li (BLOOMBERG/ 120 PARK) <[email protected]> wrote: Thanks Brian. It looks like the worker timeout. Jan 31 09:19:30 pulpp-ob-581 gunicorn[147883]: [2020-01-31 09:19:30 -0500] [147883] [CRITICAL] WORKER TIMEOUT (pid:147898) Jan 31 09:19:30 pulpp-ob-581 gunicorn[147883]: [2020-01-31 14:19:30 +0000] [147898] [INFO] Worker exiting (pid: 147898) Jan 31 09:19:31 pulpp-ob-581 gunicorn[147883]: [2020-01-31 09:19:31 -0500] [161199] [INFO] Booting worker with pid: 161199 How do I change the timeout setting in settings.py? From: [email protected] At: 01/31/20 04:18:14To: Bin Li (BLOOMBERG/ 120 PARK ) Cc: [email protected] Subject: Re: [Pulp-list] pulp 3 list packages What do the logs say about why the gunicorn process serving pulp-api is dying? Would you want to file an issue https://pulp.plan.io/issues/new so we can do some testing? As an aside, I recommend using paging when pulling so many items from an API. You could decompose your large request to more, smaller requests like: http GET localhost/pulp/api/v3/content/rpm/packages/ offset=0 limit==10000 repository_version==/pulp/api/v3/repositories/rpm/rpm/2f46d319-7997-4e86-b159-8babee4aba19/versions/1/ --timeout=200 http GET localhost/pulp/api/v3/content/rpm/packages/ offset=10000 limit==10000 repository_version==/pulp/api/v3/repositories/rpm/rpm/2f46d319-7997-4e86-b159-8babee4aba19/versions/1/ --timeout=200 What's interesting about more, smaller requests is you can likely get the data out of Pulp a lot faster since you can engage more gunicorn processes in parallel. Conceptually one large query is attractive though, so maybe we could improve that if you file it. Another idea is to limit which fields are being returned to get at the data you need faster. All the best, Brian On Thu, Jan 30, 2020 at 2:46 PM Bin Li (BLOOMBERG/ 120 PARK) <[email protected]> wrote: The rhel 7 servers rpm repo has more than 26k packages. I got an "502 Bad Gateway" error if I tried to list all of them http GET localhost/pulp/api/v3/content/rpm/packages/ limit==20000 repository_version==/pulp/api/v3/repositories/rpm/rpm/2f46d319-7997-4e86-b159-8babee4aba19/versions/1/ --timeout=200 What could cause this? Is there a fix? Thanks_______________________________________________ Pulp-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/pulp-list _______________________________________________ Pulp-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/pulp-list _______________________________________________ Pulp-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/pulp-list -- Mike DePaulo He / Him / His Service Reliability Engineer, Pulp Red Hat IM: mikedep333 GPG: 51745404
_______________________________________________ Pulp-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/pulp-list
