Horton

Re: Synchronizing slots from primary to standby

Nisha Moond <nisha.moond412@gmail.com>

View thread View on postgr.es
From: Nisha Moond <nisha.moond412@gmail.com>
To: shveta malik <shveta.malik@gmail.com>
Cc: Amit Kapila <amit.kapila16@gmail.com>, Peter Smith <smithpb2250@gmail.com>, Ajin Cherian <itsajin@gmail.com>, "Zhijie Hou (Fujitsu)" <houzj.fnst@fujitsu.com>, Masahiko Sawada <sawada.mshk@gmail.com>, Bertrand Drouvot <bertranddrouvot.pg@gmail.com>, Dilip Kumar <dilipbalaut@gmail.com>, "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>, Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>, Peter Eisentraut <peter.eisentraut@enterprisedb.com>, Bruce Momjian <bruce@momjian.us>, Ashutosh Sharma <ashu.coek88@gmail.com>, Andres Freund <andres@anarazel.de>, pgsql-hackers <pgsql-hackers@postgresql.org>, Alvaro Herrera <alvherre@alvh.no-ip.org>
Date: 2024-02-08T06:14:05Z
Lists: pgsql-hackers 
We conducted stress testing for the patch with a setup of one primary
node with 100 tables and five subscribers, each having 20
subscriptions. Then created three physical standbys syncing the
logical replication slots from the primary node.
All 100 slots were successfully synced on all three standbys. We then
ran the load and monitored LSN convergence using the prescribed SQL
checks.
Once the standbys were failover-ready, we were able to successfully
promote one of the standbys and all the subscribers seamlessly
migrated to the new primary node.

We replicated the tests with 200 tables, creating 200 logical
replication slots. With the increased load, all the tests were
completed successfully.

Minor errors (not due to patch) observed during tests -

1) When the load was run, on subscribers, the logical replication
apply workers started failing due to timeout. This is not related to
the patch as it happened due to the small "wal_receiver_timeout"
setting w.r.t. the load. To confirm, we ran the same load without the
patch too, and the same failure happened.
2) There was a buffer overflow exception on the primary node with the
'200 replication slots' case. It was not related to the patch as it
was due to short memory configuration.

All the tests were done on Windows as well as Linux environments.
Thank you Ajin for the stress test and analysis on Linux.

Commits

  1. Doc: Add the new section "Logical Replication Failover".

  2. Fix the review comments and a bug in the slot sync code.

  3. Fix a test in failover slots regression test.

  4. Fix the intermittent buildfarm failures in 040_standby_failover_slots_sync.

  5. Split XLogCtl->LogwrtResult into separate struct members

  6. Ensure that the sync slots reach a consistent state after promotion without losing data.

  7. Introduce a new GUC 'standby_slot_names'.

  8. Fix BF failure introduced by commit b3f6b14cf4.

  9. Fixups for commit 93db6cbda0.

  10. Fix BF failure in commit 93db6cbda0.

  11. Add a new slot sync worker to synchronize logical slots.

  12. Improve ERROR/LOG messages added by commits ddd5f4f54a and 7a424ece48.

  13. Disable autovacuum on primary in 040_standby_failover_slots_sync test.

  14. Fix the incorrect format specifier used in commit 7a424ece48.

  15. Change the LOG level in 040_standby_failover_slots_sync.pl to DEBUG2.

  16. Add more LOG and DEBUG messages for slot synchronization.

  17. Another try to fix BF failure introduced in commit ddd5f4f54a.

  18. Fix BF introduced in commit ddd5f4f54a.

  19. Add a slot synchronization function.

  20. Clean-ups for 776621a5e4 and 7329240437.

  21. Enhance libpqrcv APIs to support slot synchronization.

  22. Add a failover option to subscriptions.

  23. Allow setting failover property in the replication command.

  24. Allow to enable failover property for replication slots via SQL API.

  25. Add macros for looping through a List without a ListCell.

  26. Track conflict_reason in pg_replication_slots.

  27. Allow upgrades to preserve the full subscription's state.

  28. Add support for incremental backup.

  29. Simplify some logic in CreateReplicationSlot()

  30. Fix uninitialized access to InitialRunningXacts during decoding after ERROR.

  31. Flexible options for CREATE_REPLICATION_SLOT.

  32. Empty search_path in logical replication apply worker and walsender.

  33. Fix oldest xmin and LSN computation across repslots after advancing

  34. Require C99 (and thus MSCV 2013 upwards).