Re: FSM Corruption (was: Could not read block at end of the relation)

From: Ronan Dunklau <ronan(dot)dunklau(at)aiven(dot)io>
To: Noah Misch <noah(at)leadboat(dot)com>, Michael Paquier <michael(at)paquier(dot)xyz>, Patrick Stählin <patrick(dot)staehlin(at)aiven(dot)io>
Cc: pgsql-bugs <pgsql-bugs(at)lists(dot)postgresql(dot)org>
Subject: Re: FSM Corruption (was: Could not read block at end of the relation)
Date: 2024-03-06 11:08:38
Message-ID: 1979116.usQuhbGJ8B@aivenlaptop
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Le mercredi 6 mars 2024, 11:59:43 CET Patrick Stählin a écrit :
> On 3/6/24 10:31, Ronan Dunklau wrote:
> > Le mardi 5 mars 2024, 00:05:03 CET Noah Misch a écrit :
> >> I would guess this one is more risky from a performance perspective,
> >> since
> >> we'd be adding to a hotter path under RelationGetBufferForTuple().
> >> Still,
> >> it's likely fine.
> >
> > I ended up implementing this in the attached patch. The idea is that we
> > detect if the FSM returns a page past the end of the relation, and ignore
> > it. In that case we will fallback through the extension mechanism.
>
> @@ -582,7 +583,17 @@ RelationGetBufferForTuple(Relation relation, Size len,
> * We have no cached target page, so ask the FSM for an
initial
> * target.
> */
> + nblocks = RelationGetNumberOfBlocks(relation);
> targetBlock = GetPageWithFreeSpace(relation,
targetFreeSpace);
>
> I'd move the fetching of nblocks to after getting the targetBlock. This
> avoids extending the relation in the unlikely event of a FSM/relation
> extension in-between those two calls.

Good catch. Please find attached v2 of the patch.

--
Ronan Dunklau

Attachment Content-Type Size
0002-Detect-invalid-FSM-when-finding-a-block.patch text/x-patch 2.3 KB

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Tender Wang 2024-03-06 12:07:02 Re: BUG #18314: PARALLEL UNSAFE function does not prevent parallel index build
Previous Message Patrick Stählin 2024-03-06 10:59:43 Re: FSM Corruption (was: Could not read block at end of the relation)