Horton

Re: BUG #17103: WAL segments are not removed after exceeding max_slot_wal_keep_size

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Kyotaro Horiguchi <horikyota.ntt@gmail.com>, jeff.janes@gmail.com, mk@071.ovh, pgsql-bugs@lists.postgresql.org
Date: 2021-07-28T15:38:28Z
Lists: pgsql-bugs 
Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
> On 2021-Jul-19, Kyotaro Horiguchi wrote:
>> What do you think about the attached?

> Sounds sensible -- I verified the logs for one of the cases that failed
> in the buildfarm, and indeed the "checkpoint ended" message appears
> after the s2 slot is created, so it should fix the problem.  (I didn't
> actually try to reproduce the problem locally, so I didn't verify the
> fix any further than ensuring the test still passes.)

This test is still unstable :-(

https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=tern&dt=2021-07-20%2012%3A46%3A11
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=kittiwake&dt=2021-07-20%2015%3A05%3A39
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=kittiwake&dt=2021-07-28%2014%3A33%3A01
https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=kittiwake&dt=2021-07-28%2014%3A33%3A01

These all look like

#   Failed test 'check that segments have been removed'
#   at t/019_replslot_limit.pl line 226.
#          got: '000000010000000000000020'
#     expected: '000000010000000000000024'
# Looks like you failed 1 test of 16.

with varying values mentioned.  It looks to me like WAL file cleanup
is not as synchronous with slot creation as the test thinks.
Maybe it needs to loop until the oldest WAL file matches what it
expects?

			regards, tom lane

Commits

  1. Close yet another race condition in replication slot test code

  2. Make new replication slot test code even less racy

  3. Make new replication slot test code less racy

  4. Advance old-segment horizon properly after slot invalidation