Re: BUG #18499: Reindexing spgist index concurrently triggers Assert("TransactionIdIsValid(state->myXid)")

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Michael Paquier <michael(at)paquier(dot)xyz>
Cc: exclusion(at)gmail(dot)com, pgsql-bugs(at)lists(dot)postgresql(dot)org
Subject: Re: BUG #18499: Reindexing spgist index concurrently triggers Assert("TransactionIdIsValid(state->myXid)")
Date: 2024-06-16 23:30:46
Message-ID: 3050563.1718580646@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Michael Paquier <michael(at)paquier(dot)xyz> writes:
> On Sun, Jun 16, 2024 at 06:52:52PM -0400, Tom Lane wrote:
>> ... not sure there's a nice way for spginsert() to know whether it's
>> being invoked in REINDEX CONCURRENTLY or a normal INSERT/UPDATE
>> query. Can we trust indexInfo->ii_Concurrent for that?

> I am not sure to understand the redirection part for spgist, but
> except if I am missing something, we already rely on ii_Concurrent for
> other index AMs like BRIN paths to check if we are dealing with a
> concurrent build path or not. index_concurrently_build() is used
> by both CIC and REINDEX CONCURRENTLY, where the flag is set after a
> BuildIndexInfo().

Right. I was thinking that CIC wouldn't reach spginsert(), rather
spgbuild(), but it does feel a bit rickety. A separate flag would
be better.

On the whole I'm inclined to stick with the patch as I have it.
Maybe somebody would like to investigate this idea as an improvement
for v18 or later.

regards, tom lane

In response to

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Michael Paquier 2024-06-17 05:29:05 Re: BUG #18483: Segmentation fault in tests modules
Previous Message Michael Paquier 2024-06-16 23:28:10 Re: error "can only drop stats once" brings down database