Sean Chittenden <sean(at)chittenden(dot)org> writes:
> While the master was still being setup and data was being loaded, the
> system ran out of open file descriptors and Pg crashed a few times
> until this was realized and corrected.
Could you provide a bit more detail than that? Usually the behavior we
see under out-of-FDs conditions is that everything else on the system
starts to fall over badly, but PG keeps soldiering along to some extent.
I would not expect data corruption to result, anyway.
If you've got the postmaster logs from this episode, it would be useful
to see what complaints got logged.
regards, tom lane