Martijn van Oosterhout <kleptog@svana.org> writes:
> On Wed, Sep 26, 2007 at 10:51:43AM +0200, Romain Roure wrote:
>> After checking through the logs, it doesn't appear to be a problem
>> resulting from wrap-around OID's. Though the logs mention
>> transaction-wraparound may have happened.
> Please shouw us the xmin,xmax columns to the pg_class tables. But if
> you've wrapped around so far to get old column, then you passed the
> wraparound horizon 2 billion transactions ago. Please show us exactly
> what the logs say:
If there are complaints like that in the logs, it seems hugely
optimistic to suppose that you don't have a wraparound problem ...
If it is wraparound, it's possible that a VACUUM on pg_class would fix
this. It will certainly not do any harm to try it.
>> We're running PostgreSQL 8.0.1. Any help would be appreciated.
> You need to be running VACUUM...
Not to mention running a much newer release. 8.0 is still supported,
but it's up to 8.0.14 now. You're not going to find a lot of sympathy
if this turns out to have been caused by a bug that was fixed since
8.0.1 --- that subrelease was obsoleted over two years ago.
regards, tom lane