Horton

Re: Is this a problem in GenericXLogFinish()?

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Amit Kapila <amit.kapila16@gmail.com>
Cc: "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>, Alexander Lakhin <exclusion@gmail.com>, Robert Haas <robertmhaas@gmail.com>, Jeff Davis <pgsql@j-davis.com>, Heikki Linnakangas <hlinnaka@iki.fi>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2024-04-11T00:23:34Z
Lists: pgsql-hackers 
On Wed, Apr 10, 2024 at 03:28:22PM +0530, Amit Kapila wrote:
> I can understand this comment as I am aware of this code but not sure
> it would be equally easy for the people first time looking at this
> code. One may try to find the equivalent assertion in
> _hash_freeovflpage(). The alternative could be: "Ensure that the
> required flags are set when there are no tuples. See
> _hash_freeovflpage().". I am also fine if you prefer to go with your
> proposed comment.

Yes, I can see your point about why that's confusing.  Your suggestion
is much better, so after a second look I've used your version of the
comment and applied the patch on HEAD.

I am wondering if we have other problems like that with dirty buffers
at replay.  Perhaps I should put my nose more onto the replay paths
and extend these automated checks with wal_consistency_checking.
--
Michael

Commits

  1. Fix inconsistency with replay of hash squeeze record for clean buffers

  2. Set LSN for wbuf in _hash_freeovflpage() iff wbuf is modified.

  3. Fix an uninitialized access in hash_xlog_squeeze_page().

  4. Use REGBUF_NO_CHANGE at one more place in the hash index.

  5. Assert that buffers are marked dirty before XLogRegisterBuffer().

  6. Fix bug in GenericXLogFinish().

  7. Improve hash index bucket split behavior.