Horton

Re: Race condition in recovery?

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

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
Cc: robertmhaas@gmail.com, dilipbalaut@gmail.com, hlinnaka@iki.fi, pgsql-hackers@lists.postgresql.org
Date: 2021-06-11T15:23:27Z
Lists: pgsql-hackers 
Kyotaro Horiguchi <horikyota.ntt@gmail.com> writes:
>> ==~_~===-=-===~_~== pgsql.build/src/bin/pg_verifybackup/tmp_check/log/003_corruption_primary.log ==~_~===-=-===~_~==
>> ...
>> 2021-06-08 16:17:41.706 CEST [51792:9] 003_corruption.pl LOG:  received replication command: START_REPLICATION SLOT "pg_basebackup_51792" 0/B000000 TIMELINE 1
>> 2021-06-08 16:17:41.706 CEST [51792:10] 003_corruption.pl STATEMENT:  START_REPLICATION SLOT "pg_basebackup_51792" 0/B000000 TIMELINE 1
>> (log ends here)

> There seems like some hardware failure?

conchuela has definitely evinced flakiness before.  Not sure what's
up with it, but I have no problem with writing off non-repeatable
failures from that machine.  In any case, it's now passed half a
dozen times in a row on HEAD, so I think we can say that it's okay
with this test.  That leaves jacana, which I'm betting has a
Windows portability issue with the new test.

			regards, tom lane

Commits

  1. Further refinement of stuck_on_old_timeline recovery test

  2. Adjust new test case to set wal_keep_segments.

  3. Adjust new test case to set wal_keep_size.

  4. Fix corner case failure of new standby to follow new primary.

  5. Back-port a few PostgresNode.pm methods.

  6. Allow PostgresNode.pm's backup method to accept backup_options.

  7. Add end-to-end testing of pg_basebackup's tar-format output.

  8. Allow PostgresNode.pm tests to wait for catchup

  9. Delay reading timeline history file until it's fetched from master.