> > My guess is the internal structure of one of the tuples is corrupt.
>
> So it would seem. Evidently there's at least one more corrupted page
> besides the ones you were able to identify before.
I punted last night and reloaded the customer table from a backup. Turns
out I could recreate the most critical records from other sources. Life
goes on.
I would still like to refine my pganal tool to look inside of tuples. Where
should I look to find information about the internal structure? Is it
parsable at some level on its own or do I have to consult the system tables
to determine stucture. I suspect this might get more complicated once TOAST
is available.
Bryan White