| From: | Andrew Dunstan <andrew(at)dunslane(dot)net> |
|---|---|
| To: | Andres Freund <andres(at)anarazel(dot)de> |
| Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Tomas Vondra <tomas(dot)vondra(at)enterprisedb(dot)com>, Nikita Glukhov <n(dot)gluhov(at)postgrespro(dot)ru>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: ltree_gist indexes broken after pg_upgrade from 12 to 13 |
| Date: | 2022-03-10 15:43:06 |
| Message-ID: | ba6da883-ad96-6701-23c0-e6b7c8b7dfc5@dunslane.net |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 3/6/22 17:33, Andres Freund wrote:
> Hi,
>
> On 2022-03-06 07:46:04 -0500, Andrew Dunstan wrote:
>> This is an area not currently touched by the buildfarm's cross version
>> upgrade testing, which basically compares a pre-upgrade and post-upgrade
>> dump of the databases. The upgraded cluster does contain
>> contrib_regression_ltree.
>>
>> I'm open to suggestions on how we might improve the buildfarm's testing
>> of upgraded indexes generally.
> One thing that's likely worth doing as part of the cross version upgrade test,
> even if it wouldn't even help in this case, is to run amcheck post
> upgrade. Just dumping data isn't going to touch indices at all.
>
> A sequence of
> pg_upgrade; amcheck; upgrade all extensions; amcheck;
> would make sense.
>
See
<https://github.com/PGBuildFarm/client-code/commit/191df23bd25eb5546b0989d71ae92747151f9f39>
This will be in the next release
cheers
andrew
--
Andrew Dunstan
EDB: https://www.enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Simon Riggs | 2022-03-10 15:48:54 | Re: support for MERGE |
| Previous Message | Tom Lane | 2022-03-10 15:39:39 | Re: ltree_gist indexes broken after pg_upgrade from 12 to 13 |