Re: PostgresSQL 9.5.21 very slow to connect and perform basic queries
bruno da silva <brunogiovs@gmail.com>
From: bruno da silva <brunogiovs@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-performance@lists.postgresql.org
Date: 2022-08-02T17:58:02Z
Lists: pgsql-performance
Hello. Are you quite sure this is a 9.5.21 version of the pg_stat_statements extension? *I got version 1.3 from SELECT * FROM pg_extension;* Is it possible that the pg_stat_tmp directory has been made non-writable? *hard to tell if it was made non-writable during the outage. but now it is writable.* Thanks On Tue, Aug 2, 2022 at 1:25 PM Tom Lane <tgl@sss.pgh.pa.us> wrote: > bruno da silva <brunogiovs@gmail.com> writes: > > Do you have a lot of especially long statements being tracked > > in the pg_stat_statements view?* well, the view was showing the query > > column null.* > > * but looking on pgss_query_texts.stat there are very large sql > > statements, of around ~ 400kb, multiple thousands. * > > Hm. We try to recover from such failures by (a) resetting all the view's > query text fields to null and (b) truncating the file --- well, unlinking > it and creating it as empty. It seems like (a) happened and (b) didn't. > It's pretty hard to explain that from the code though. Are you quite > sure this is a 9.5.21 version of the pg_stat_statements extension? > Is it possible that the pg_stat_tmp directory has been made non-writable? > > regards, tom lane > > > > > > Are there any other signs of distress > > in the postmaster log, like complaints about being unable to write > > pgss_query_texts.stat? *no, just complaints for reading it. * > > > Thanks > > > On Tue, Aug 2, 2022 at 11:59 AM Tom Lane <tgl@sss.pgh.pa.us> wrote: > > >> bruno da silva <brunogiovs@gmail.com> writes: > > After more investigation, we found that pgss_query_texts.stat of a size > >> of > > 2.2GB. and this deployment has a 32bit pg. > >> > >> Hm ... we've heard one previous report of pg_stat_statements' query text > >> file getting unreasonably large, but it's not clear how that can come > >> to be. Do you have a lot of especially long statements being tracked > >> in the pg_stat_statements view? Are there any other signs of distress > >> in the postmaster log, like complaints about being unable to write > >> pgss_query_texts.stat? > >> > >> regards, tom lane > >> > > > > -- > > Bruno da Silva > > -- Bruno da Silva
Commits
-
Be more wary about 32-bit integer overflow in pg_stat_statements.
- c67c2e2a2939 16.0 landed
- dd414bf4e047 10.22 landed
- 82ebc70d1c7f 15.0 landed
- 6b67db10c366 13.8 landed
- 6608a4305636 12.12 landed
- 17fd203b414e 14.5 landed
- 06f6a07ba465 11.17 landed