Horton

Re: [BUG] non archived WAL removed during production crash recovery

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Andres Freund <andres@anarazel.de>
Cc: Jehan-Guillaume de Rorthais <jgdr@dalibo.com>, Kyotaro Horiguchi <horikyota.ntt@gmail.com>, masao.fujii@oss.nttdata.com, pgsql-bugs@lists.postgresql.org
Date: 2020-04-24T07:05:10Z
Lists: pgsql-bugs, pgsql-hackers 
On Thu, Apr 23, 2020 at 06:48:56PM -0700, Andres Freund wrote:
> But don't we pretty much know this already from the state of the system?
> During crash recovery there's nothing running RemoveOldXLogFiles() but
> the startup process. Right? And in DB_IN_ARCHIVE_RECOVERY it should only
> happen as part of restartpoints (i.e. the checkpointer).
> 
> Did you add the new shared state to avoid deducing things from the
> "environment"? If so, it should really be mentioned in the commit
> message & code. Because:

Hmm.  Sorry, I see your point.  The key of the logic here is from
XLogArchiveCheckDone() which could be called from other processes than
the startup process.  There is one code path at the end of a base
backup for backup history files where not using a shared state would
be a problem.
--
Michael

Commits

  1. Fix interpolation in test name.

  2. Add more TAP coverage for archive status with crash recovery of standbys

  3. Fix handling of WAL segments ready to be archived during crash recovery

  4. Fix WAL recycling on standbys depending on archive_mode