Re: hash_xlog_split_allocate_page: failed to acquire cleanup lock
Robert Haas <robertmhaas@gmail.com>
From: Robert Haas <robertmhaas@gmail.com>
To: Amit Kapila <amit.kapila16@gmail.com>
Cc: Andres Freund <andres@anarazel.de>, Tom Lane <tgl@sss.pgh.pa.us>, Mark Dilger <mark.dilger@enterprisedb.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>,
Thomas Munro <thomas.munro@gmail.com>
Date: 2022-10-31T17:10:40Z
Lists: pgsql-hackers
On Mon, Oct 31, 2022 at 7:27 AM Amit Kapila <amit.kapila16@gmail.com> wrote: > Agreed, I think the important point to decide is what to do for > back-branches. We have the next minor release in a few days' time and > this is the last release for v10. I see the following options based on > the discussion here. > > a. Use the code change in 0001 from email [1] and a comment change > proposed by Robert in email [2] to fix the bug reported. This should > be backpatched till v10. Then separately, we can consider committing > something like 0002 from email [1] as a HEAD-only patch. > b. Use the code change in 0001 from email [1] to fix the bug reported. > This should be backpatched till v10. Then separately, we can consider > committing something like 0002 from email [1] as a HEAD-only patch. > c. Combine 0001 and 0002 from the email [1] and push them in all > branches till v10. > > I prefer going with (a). I vote for (a) or (b) for now, and we can consider what else to do later. It might even include back-patching. But fixing things that are causing problems we can see seems to me to have higher priority than fixing things that are not. -- Robert Haas EDB: http://www.enterprisedb.com
Commits
-
Fix cleanup lock acquisition in SPLIT_ALLOCATE_PAGE replay.
- e848be60b5cf 16.0 landed
- e49e191815b6 15.2 landed
- 9693f190076e 14.7 landed
- 20c223336301 13.10 landed
- 4dccccb37e0b 12.14 landed
- 1703033f896a 11.19 landed