Horton

RE: test_decoding assertion failure for the loss of top-sub transaction relationship

Hayato Kuroda (Fujitsu) <kuroda.hayato@fujitsu.com>

View thread View on postgr.es
From: "kuroda.hayato@fujitsu.com" <kuroda.hayato@fujitsu.com>
To: 'Amit Kapila' <amit.kapila16@gmail.com>
Cc: Kyotaro Horiguchi <horikyota.ntt@gmail.com>, "dilipbalaut@gmail.com" <dilipbalaut@gmail.com>, "osumi.takamichi@fujitsu.com" <osumi.takamichi@fujitsu.com>, "pgsql-hackers@lists.postgresql.org" <pgsql-hackers@lists.postgresql.org>
Date: 2022-09-07T02:06:02Z
Lists: pgsql-hackers

Attachments

Dear Amit,

Thanks for giving comments!

> Did you get this new assertion failure after you applied the patch for
> the first failure? Because otherwise, how can you reach it with the
> same test case?

The first failure is occurred only in the HEAD, so I did not applied the first patch
to REL14 and REL15.
This difference is caused because the commit [Fix catalog lookup...] in REL15(272248a) and older is different
from the HEAD one.
In order versions SnapBuildXidSetCatalogChanges() has been added. In the function
a transaction will be marked as containing catalog changes if the transaction is in InitialRunningXacts,
and after that the relation between sub-top transactions is assigned based on the parsed->subxact.
The marking avoids the first failure, but the assignment triggers new failure.


> About patch:
> else if (sub_needs_timetravel)
>   {
> - /* track toplevel txn as well, subxact alone isn't meaningful */
> + elog(DEBUG2, "forced transaction %u to do timetravel due to one of
> its subtransaction",
> + xid);
> + needs_timetravel = true;
>   SnapBuildAddCommittedTxn(builder, xid);
> 
> Why did you remove the above comment? I think it still makes sense to retain it.

Fixed.

Best Regards,
Hayato Kuroda
FUJITSU LIMITED

Commits

  1. Fix executing invalidation messages generated by subtransactions during decoding.

  2. Fix assertion failures while processing NEW_CID record in logical decoding.