Re: Necessary actions after an OS upgrade

On Tue, Sep 30, 2025 at 5:50 AM Zwettler Markus (OIZ) <
markus(dot)zwettler(at)zuerich(dot)ch> wrote:

> We use Crunchy Postgres for Kubernetes (CPK). With CPK 5.8, there will be
> an OS upgrade from UBI8 to UBI9 images. As described in both the Crunchy
> and Postgres documentation, conflicts due to collation version mismatches
> following an OS upgrade can lead to incorrect queries.
>
>
>
>
> https://access.crunchydata.com/documentation/postgres-operator/latest/tutorials/cluster-management/update-cluster#changing-base-images
>
>
> https://www.postgresql.org/docs/current/sql-altercollation.html#SQL-ALTERCOLLATION-NOTES
>
>
>
> According to the documentation, we plan to perform a “reindex database”
>

I would only rebuild text indices. Most indices are (probably) INTEGER or
BIGINT, so this would save you a lot of time

> and a “alter database refresh collation version” immediately after the
> upgrade.
>
>
>
>
>
> Is there anything else to consider? I'm wondering if this is sufficient.
>
>
>
> What about materialized views,
>

> (text) function-based indexes
>

That would be caught by REINDEX ALL, no?

> , and extensions? Couldn't problems arise here as well? None of the
> documentation addresses this. Am I thinking about this incorrectly?
>
>
>
> Ideally, I would like to have a precise list of everything that needs to
> be considered and done afterwards.
>
>
--
Death to <Redacted>, and butter sauce.
Don't boil me, I'm still alive.
<Redacted> lobster!