Hey Tom,
Had a small query here.
If we have streaming replication enabled, and if we have corruption like this due to power loss etc,
is it safe to assume we failover to standby and we should be good (atleast in most cases)
I wanted to understand if the system catalog is corrupted, for some reason, will it/will it not stream replicate the
corrupteddata to standby.
Or they are a property of an individual database server and will be rebuilt/updated on pg_dump/restore.
I am sorry if I am diverting the Drako's query or if that is not related, I'll ask another question then __
But I guess maybe the answer may help Drako too.
Thanks,
Vijay
On 8/20/18, 12:53 AM, "Tom Lane" <tgl@sss.pgh.pa.us> wrote:
DrakoRod <drakoflames@hotmail.com> writes:
> I reindex the pg_proc table and reindex finished correctly but I try read
> the tables or make the dump and same error, reindexed the database and show
> me this error:
> ERROR: cache lookup failed for function 125999
This suggests you've actually lost some entries from pg_proc. Have
you had any system crashes or suchlike?
> Any suggestions, If I run pg_upgrade to 10, will these errors be corrected?
No, pg_upgrade can't magically restore data that's not there.
regards, tom lane