Hi Tom,
thanks for your reply
> You could try reindexing the system tables, but I don't have a lot of
> hope for that. See the REINDEX man page.
I allready tried this ... didn't work out.
OK - I am sorry, but the ECC-RAM I ordered didn't run in the old server
so I had to order a new server as fast as possible... it's going to run
in a about 20-30 minutes hopefully....
How can I start postgres in stand-alone mode with "-o -P" as you wrote
and then do a pg_dump ? Or did I get you wrong?
> > or is only an initb possible?
>
> I'd recommend an initdb in any case on the new server. Who knows how
> much corruption has snuck into your existing database files? You
> certainly cannot copy them over to the new server without great risk.
> If you are able to do a pg_dump then you could use the dump file with
> some amount of assurance.
>
> This is iffy, but if REINDEX fails, you might try starting the
> postmaster with "-o -P" (disable use of system indexes) and then
> see if you can do a pg_dump. I think that might get around the loss
> of pg_statistic_relid_att_index.
>
> In any case it's foolish to keep running a live application on such
> clearly broken hardware. The longer you run, the more corrupt your
> database will get. I suggest that you do nothing except an immediate
> pg_dump...
>
> regards, tom lane