Horton

Re: BUG: Postgres 14 + vacuum_defer_cleanup_age + FOR UPDATE + UPDATE

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Peter Eisentraut <peter.eisentraut@enterprisedb.com>
Cc: Mark Dilger <mark.dilger@enterprisedb.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2023-05-10T18:04:08Z
Lists: pgsql-hackers 
Hi,

On 2023-05-10 17:44:07 +0200, Peter Eisentraut wrote:
> On 12.03.23 00:41, Andres Freund wrote:
> > Hi,
> > 
> > On 2023-03-11 15:34:55 -0800, Mark Dilger wrote:
> > > > On Mar 11, 2023, at 3:22 PM, Andres Freund <andres@anarazel.de> wrote:
> > > > 
> > > > Something like the attached.
> > > 
> > > I like that your patch doesn't make the test longer.  I assume you've already run the tests and that it works.
> > 
> > I did check that, yes :). My process of writing perl is certainly, uh,
> > iterative. No way I would get anything close to working without testing it.
> > 
> > CI now finished the tests as well:
> > https://cirrus-ci.com/build/6675457702100992
> > 
> > So I'll go ahead and push that.
> 
> There is a small issue with this commit (a4f23f9b3c).
> 
> In src/bin/pg_amcheck/t/004_verify_heapam.pl, there is code to detect
> whether the page layout matches expectations and if not it calls plan
> skip_all.

Some of these skip_all's don't seem like a good idea. Why is a broken
relfrozenxid a cause for skipping a test? But anyway, that's really unrelated
to the topic at hand.


> This commit adds a test
> 
> is(scalar @lp_off, $ROWCOUNT, "acquired row offsets");
> 
> *before* that skip_all call.  This appears to be invalid.  If the skip_all
> happens, you get a complaint like
> 
> t/004_verify_heapam.pl (Wstat: 0 Tests: 1 Failed: 0)
>   Parse errors: Bad plan.  You planned 0 tests but ran 1.
> 
> We could move the is() test after all the skip_all's.  Any thoughts?

I think the easiest fix is to just die if we can't get the offsets - it's not
like we can really continue afterwards...

Greetings,

Andres Freund

Commits

  1. Fix incorrect TAP test ordering

  2. pg_amcheck: Minor test speedups

  3. amcheck: Fix FullTransactionIdFromXidAndCtx() for xids before epoch 0

  4. amcheck: Fix ordering bug in update_cached_xid_range()

  5. Fix corruption due to vacuum_defer_cleanup_age underflowing 64bit xids

  6. Add hardening to catch invalid TIDs in indexes.

  7. Use full 64-bit XID for checking if a deleted GiST page is old enough.

  8. Avoid early reuse of btree pages, causing incorrect query results.