Thread: pgsql: Don't ERROR on PreallocXlogFiles() race condition.

pgsql: Don't ERROR on PreallocXlogFiles() race condition.

From
Michael Paquier
Date:
Don't ERROR on PreallocXlogFiles() race condition.

Before a restartpoint finishes PreallocXlogFiles(), a startup process
KeepFileRestoredFromArchive() call can unlink the preallocated segment.
If a CHECKPOINT sql command had elicited the restartpoint experiencing
the race condition, that sql command failed.  Moreover, the restartpoint
omitted its log_checkpoints message and some inessential resource
reclamation.  Prevent the ERROR by skipping open() of the segment.
Since these consequences are so minor, no back-patch.

This commit has been applied as of 2b3e4672f760 in v15 and newer
versions.  This is required on stable branches of v13 and v14 to fix a
regression reported by Noah Misch, introduced by 1f95181b44c8, causing
spurious failures in archive recovery (neither streaming nor archive
recovery) with concurrent restartpoints.  The backpatched versions of
the patches have been aligned on these branches by me, Noah Misch is the
author.  Tests have been conducted by the both of us.

Reported-by: Arun Thirupathi
Author: Noah Misch <noah@leadboat.com>
Discussion: https://postgr.es/m/20210202151416.GB3304930@rfd.leadboat.com
Discussion: https://postgr.es/m/20250306193013.36.nmisch@google.com
Backpatch-through: 13

Branch
------
REL_13_STABLE

Details
-------
https://git.postgresql.org/pg/commitdiff/cbed472a93417abc0411b5c8a0ac4d20ec7f4911
Author: Noah Misch <noah@leadboat.com>

Modified Files
--------------
src/backend/access/transam/xlog.c     | 79 +++++++++++++++++++++++++----------
src/backend/replication/walreceiver.c |  4 +-
src/include/access/xlog.h             |  2 +-
3 files changed, 58 insertions(+), 27 deletions(-)