Horton

Re: [PoC] pg_upgrade: allow to upgrade publisher node

Amit Kapila <amit.kapila16@gmail.com>

View thread View on postgr.es
From: Amit Kapila <amit.kapila16@gmail.com>
To: "Hayato Kuroda (Fujitsu)" <kuroda.hayato@fujitsu.com>
Cc: Julien Rouhaud <rjuju123@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2023-07-17T12:49:44Z
Lists: pgsql-hackers 
On Fri, Jun 30, 2023 at 7:29 PM Hayato Kuroda (Fujitsu)
<kuroda.hayato@fujitsu.com> wrote:
>
> I have analyzed more, and concluded that there are no difference between manual
> and shutdown checkpoint.
>
> The difference was whether the CHECKPOINT record has been decoded or not.
> The overall workflow of this test was:
>
> 1. do INSERT
> (2. do CHECKPOINT)
> (3. decode CHECKPOINT record)
> 4. receive feedback message from standby
> 5. do shutdown CHECKPOINT
>
> At step 3, the walsender decoded that WAL and set candidate_xmin_lsn. The stucktrace was:
> standby_decode()->SnapBuildProcessRunningXacts()->LogicalIncreaseXminForSlot().
>
> At step 4, the confirmed_flush of the slot was updated, but ReplicationSlotSave()
> was executed only when the slot->candidate_xmin_lsn had valid lsn. If step 2 and
> 3 are misssed, the dirty flag is not set and the change is still on the memory.
>
> FInally, the CHECKPOINT was executed at step 5. If step 2 and 3 are misssed and
> the patch from Julien is not applied, the updated value will be discarded. This
> is what I observed. The patch forces to save the logical slot at the shutdown
> checkpoint, so the confirmed_lsn is save to disk at step 5.
>

I see your point but there are comments in walsender.c which indicates
that we also wait for step-5 to get replicated. See [1] and comments
atop walsender.c. If this is true then we don't need a special check
as you have in patch 0003 or at least it doesn't seem to be required
in all cases.

[1] -
/*
* When SIGUSR2 arrives, we send any outstanding logs up to the
* shutdown checkpoint record (i.e., the latest record), wait for
* them to be replicated to the standby, and exit. ...
*/

-- 
With Regards,
Amit Kapila.

Commits

  1. Fix issues in binary_upgrade_logical_slot_has_caught_up().

  2. Fix a random failure in 003_logical_slots.pl.

  3. Fix a test in 003_logical_slots.

  4. Fix uninitialized slot array access during the upgrade.

  5. Fix the test 003_logical_slots.

  6. Commit b195e6d482 forgot to update meson.build.

  7. Use shorter file names in the upgrade logical slots test.

  8. Migrate logical slots to the new node during an upgrade.

  9. Flush logical slots to disk during a shutdown checkpoint if required.

  10. Prevent possibility of panics during shutdown checkpoint.