On Mon, Nov 07, 2022 at 09:02:26AM -0500, Tom Lane wrote:
> Stefan Froehlich <postgresql@froehlich.priv.at> writes:
> > On Mon, Nov 07, 2022 at 08:17:10AM -0500, Mladen Gogala wrote:
> >> On 11/7/22 06:19, Laurenz Albe wrote:
> >>> Don't continue to work with that cluster even if everything seems OK now.
> >>> "pg_dumpall" and restore to a new cluster on good hardware.
>
> >> Why would that be necessary if the original machine works well now?
>
> > I can understand the idea not to trust hardware anymore once a (not
> > clearly identified) problem occured.
>
> > In this case new hardware would - for reasons beyond the scope of
> > this list - not be any more or less trustworthy than the existing
> > one and thus (IMO) not make any difference.
>
> Whether you want to continue to trust the hardware or not is your
> call. It'd still be recommendable to pg_dumpall and restore into
> a freshly-initdb'd cluster, because otherwise you can't be real
> sure that you identified and cleared all the data corruption.
Thanks, yes. This is in fact on my schedule for the next weekend as
it implies a downtime of serveral hours.
Bye,
Stefan
