Horton

Re: Temporary tables prevent autovacuum, leading to XID wraparound

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Alvaro Herrera <alvherre@2ndquadrant.com>, Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>, tsunakawa.takay@jp.fujitsu.com, robertmhaas@gmail.com, pgsql-hackers@postgresql.org
Date: 2018-08-08T16:42:29Z
Lists: pgsql-hackers 
On Wed, Aug 08, 2018 at 10:50:34AM -0400, Tom Lane wrote:
> Alvaro Herrera <alvherre@2ndquadrant.com> writes:
>> I'm unsure about pg11 -- is it a backbranch already or not?  Since we've
>> released beta3 already, ISTM we should consider it so.
> 
> No.  See
> https://www.postgresql.org/message-id/12725.1533737052%40sss.pgh.pa.us
> 
> At this point I'd judge it's still more valuable to minimize differences
> between v11 and v12 than to avoid ABI breakages in v11.  It's unlikely
> this would be the last ABI break before 11.0 anyway.

Thanks!  Based on that I will get this patch into HEAD and 11 pretty
soon.  v10 is definitely out of scope, even if it would have been
interesting to make autovacuum more aggressive there as well, as that's
where the last improvements have been done.
--
Michael

Commits

  1. Clarify comment about assignment and reset of temp namespace ID in MyProc

  2. Make autovacuum more aggressive to remove orphaned temp tables

  3. Don't count background workers against a user's connection limit.