On 2/8/23 23:53, Paul McGarry wrote:
> I have three databases, two of databases where I am experiencing the
> issue below.
>
> The first database was created from a dump in Feb 2022 (a few weeks
> after the time period for which I seem to have problematic indexes, maybe).
> The second database was then cloned from the first (ie filesystem level
> copy) soon after.
What where the commands for the dump/restore?
Second database, same machine or different one?
> Since then all databases have undergone a number of minor version
> upgrades, including to 13.9 and an OS update last week for the two
> problem databases (the other is still on 13.8).
The OS's and the update to?
>
> Now, a process which does clears some data > 13 months old is getting an
> error when trying to do that update.
>
> My suspicion is that either:
> - there was probably an issue with the index 12 months ago and that
> problem was copied when I cloned the database, and is just becoming
> apparent now a script is accessing 13 month olf data.
> - something in our recent upgrade has caused the problem.
A dump/restore would recreate the index at the restore.
>
> The third database is still on 13.8, and with some OS updates pending,
> and is not experiencing the problem.
I'm betting OS change as culprit.
>
> Any suggestions on what else I should look into, in particular anything
> I should check before upgrading the remaining 13.8 DB to 13.9?
>
> Thanks for any help,
>
> Paul
--
Adrian Klaver
adrian.klaver@aklaver.com