David Wall <d(dot)wall(at)computer(dot)org> writes:
> Thanks, Robert. Actually, I noted that each time I ran my backup
> script, the OID mentioned in the error was bigger than previous number.
That's fairly interesting, but can you provide a self-contained test
case? The reindexdb script really doesn't do anything except invoke
REINDEX, so I see no way to blame it.
regards, tom lane