Обсуждение: Update on
Hello
We use PG 11.9 and are encountering the problem with connection limit
and parallel workers aborting a totally valid query ("CONTEXT: parallel
worker" error)
I found there was a thread concerning this problem and even a patch was
developed:
https://www.postgresql.org/message-id/CAB7nPqQKJPgNkg0VQguUP1u%2BpDdk1qF16Qh5MtDQEtSMG4S-fg%40mail.gmail.com
but no news since 2017
Do you know the status of the patch or a modification to remove the
parallel worker from the connection count against user (also server?)
CONNECTION_LIMIT or to just not launch workers if not possible?
Without that we'll have to configure a crazy number of user
connection_limit (not avoiding the problem but limiting the probability)
or remove parallel configuration, any of which would be a pity
Thanks in advance
Best regards
Martin
--
Martin Querleu - Directeur Général
EffiSYS (www.effitrace.fr - www.logistique-e-commerce.fr)
martin.querleu@effisys.fr
3, rue Gustave Delory
59000 Lille
Tél: +33 9 54 28 38 76
Vous rencontrez un problème d'utilisation sur effitr@ce?
=====> écrivez à support@effisys.fr
Vous rencontrez un problème technique au niveau des échanges de données?
=====> écrivez à supervision@effisys.fr
On Friday, November 27, 2020, EffiSYS / Martin Querleu <martin.querleu@effisys.fr> wrote:
Hello
We use PG 11.9 and are encountering the problem with connection limit and parallel workers aborting a totally valid query ("CONTEXT: parallel worker" error)
I found there was a thread concerning this problem and even a patch was developed: https://www.postgresql.org/message-id/CAB7nPqQKJPgNkg0VQguUP 1u%2BpDdk1qF16Qh5MtDQEtSMG4S- fg%40mail.gmail.com but no news since 2017
Do you know the status of the patch or a modification to remove the parallel worker from the connection count against user (also server?) CONNECTION_LIMIT or to just not launch workers if not possible?
That thread was resolved with a committed patch [1] mid-thread, then got hijacked a bit...
David J.
Hello David
After reading the commit I understand it modifies the management so that parallel workers don't count against connection_limit so I don't understand why I experience the problem
What do you mean by "hijacked"?
Best regards
Martin
On 27/11/2020 19:30, David G. Johnston wrote:
After reading the commit I understand it modifies the management so that parallel workers don't count against connection_limit so I don't understand why I experience the problem
What do you mean by "hijacked"?
Best regards
Martin
On 27/11/2020 19:30, David G. Johnston wrote:
On Friday, November 27, 2020, EffiSYS / Martin Querleu <martin.querleu@effisys.fr> wrote:Hello
We use PG 11.9 and are encountering the problem with connection limit and parallel workers aborting a totally valid query ("CONTEXT: parallel worker" error)
I found there was a thread concerning this problem and even a patch was developed: https://www.postgresql.org/message-id/CAB7nPqQKJPgNkg0VQguUP 1u%2BpDdk1qF16Qh5MtDQEtSMG4S- fg%40mail.gmail.com but no news since 2017
Do you know the status of the patch or a modification to remove the parallel worker from the connection count against user (also server?) CONNECTION_LIMIT or to just not launch workers if not possible?That thread was resolved with a committed patch [1] mid-thread, then got hijacked a bit...David J.
-- Martin Querleu - Directeur Général EffiSYS (www.effitrace.fr - www.logistique-e-commerce.fr) martin.querleu@effisys.fr 3, rue Gustave Delory 59000 Lille Tél: +33 9 54 28 38 76 Vous rencontrez un problème d'utilisation sur effitr@ce? =====> écrivez à support@effisys.fr Vous rencontrez un problème technique au niveau des échanges de données? =====> écrivez à supervision@effisys.fr
On Mon, Nov 30, 2020 at 2:04 AM EffiSYS / Martin Querleu <martin.querleu@effisys.fr> wrote:
Hello David
After reading the commit I understand it modifies the management so that parallel workers don't count against connection_limit so I don't understand why I experience the problem
Nor do I. Hoping a -hacker responds here, I just wanted to comment about the thread/patch that you mentioned.
What do you mean by "hijacked"?
After David's patch got posted, ending the thread in theory, Peter added a slightly related patch to the thread, "hijacking it" to a new purpose. What likely happened (though I didn't look) is that the new patch was realized to have been misplaced and moved to a new thread, but no one noted that on the old thread and so it looks as if that thread went unresolved if one doesn't read closely and assume that the "I'll commit this tomorrow" actually happened - which it did but no link to the commit was added to the thread.
David J.