Bricklen Anderson <bricklen@shaw.ca> writes:
>> Once I recompile the function, I no longer get that message. Is there
>> anything else that I can check or do to make this stop happening? Or is
>> this a sign of things to come (possible corruption, etc?)
Well, the original error sounds like a disk drive lossage ... you might
want to think about replacing that drive sometime soon, before it drops
data from someplace more critical than an index.
In the meantime, look through the pg_proc.prolang column for entries
that don't match the OID of any row in pg_language. Probably you
could just UPDATE the ones that are wrong to make them match the OIDs
of the new rows. I'd suggest a dump and restore of the database
after you think you have it right, just to make sure everything
is sane and to get pg_depend back in sync.
regards, tom lane