Horton

RE: Synchronizing slots from primary to standby

Zhijie Hou (Fujitsu) <houzj.fnst@fujitsu.com>

View thread View on postgr.es
From: "Zhijie Hou (Fujitsu)" <houzj.fnst@fujitsu.com>
To: shveta malik <shveta.malik@gmail.com>, Peter Smith <smithpb2250@gmail.com>
Cc: "Drouvot, Bertrand" <bertranddrouvot.pg@gmail.com>, Amit Kapila <amit.kapila16@gmail.com>, Nisha Moond <nisha.moond412@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>, Ajin Cherian <itsajin@gmail.com>, Alvaro Herrera <alvherre@alvh.no-ip.org>
Date: 2023-12-18T12:11:22Z
Lists: pgsql-hackers 
On Monday, December 11, 2023 5:31 PM shveta malik <shveta.malik@gmail.com> wrote:
> 
> On Thu, Dec 7, 2023 at 1:33 PM Peter Smith <smithpb2250@gmail.com>
> wrote:
> >
> > Hi.
> >
> > Here are my review comments for patch v43-0002.
> >
> 
> > ======
> > src/backend/access/transam/xlogrecovery.c
> >
> > 13.
> > + /*
> > + * Shutdown the slot sync workers to prevent potential conflicts between
> > + * user processes and slotsync workers after a promotion. Additionally,
> > + * drop any slots that have initiated but not yet completed the sync
> > + * process.
> > + */
> > + ShutDownSlotSync();
> > + slotsync_drop_initiated_slots();
> > +
> >
> > Is this where maybe the 'sync_state' should also be updated for
> > everything so you are not left with confusion about different states
> > on a node that is no longer a standby node?
> >
> 
> yes, this is the place. But this needs more thought as it may cause
> too much disk activity during promotion. so let me analyze and come
> back.

Per off-list discussion with Amit.

I think it's fine to keep both READY and NONE on a primary. Because even if we
update the sync_state from READY to NONE on promotion, it doesn't reduce the
complexity for the handling of READY and NONE state. And it's not
straightforward to choose the right place to update sync_state, here is the
analysis:

(related steps on promotion)
1 (patch) shutdown slotsync worker
2 (patch) drop 'i' state slots.
3 remove standby.signal and recovery.signal
4 switch to a new timeline and write the timeline history file 
5 set SharedRecoveryState = RECOVERY_STATE_DONE which means RecoveryInProgress() will return false.

We could not update the sync_state before step 3 because if the update fails after
updating some of slots' state, then the server will be shutdown leaving some
'NONE' state slots. After restarting, the server is still a standby so the slot
sync worker will fail to sync these 'NONE' state slots.

We also could not update it after step 3 and before step 4. Because if any ERROR
when updating, then after restarting the server, although the server will
become a master(as standby.signal is removed), but it can still be made as a
active standby by creating a standby.signal file because the timeline has not
been switched. And in this case, the slot sync worker will also fail to sync
these 'NONE' state slots.

Updating the sync_state after step 4 and before step 5 is OK, but still
It doesn't simplify the handling for both READY and NONE state slots.
Therefore, I think we can retain the READY state slots after promotion as they
can provide information about the slot's origin. I added some comments around
slotsync cleanup codes (in FinishWalRecovery) to mentioned the reason.

Best Regards,
Hou zj

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).