> Not if you don't know how to run PostgreSQL or your machines, no.
> Migrating to a new platform without sufficient testing or expertise
> is always unwise (sorry, but you deserved that for the dig at
> Postgres).
Heh. Well it's most certainly the case I'm still learning about Postgres
(aren't we all?), but I did run it for a few months on unimportant stuff and
to be honest, this application I'm having trouble with at the moment is not
real important even though it's production. The database is being
continuously populated and drained by different applications (its a spool
area mainly), so activity is high even though volume is not.
> Either:
> 1. You have a bad PostgreSQL install. What OS are you using?
Debian Woody with 2.4.18 Linux kernel. Postgres install from apt-get
(7.2.1).
> 2. You are out of disk space on the PostgreSQL partition.
Approx 15gig free.
> 3. You are having hardware errors, such as a corrupt HDD.
No other untoward signs.
> 4. Your OS is having disk or file management errors.
No other untoward signs.
> Regardless, you'd better shutdown Postgres and defer all work on the
> database until you've diagnosed your hardware/configuration problem. I
> hope you have a previous backup, as you may find that you need to
> revert to an earlier backup to restore your data.
In this instance, simply dropping the index and recreating it solved the
problem. I'll continue using this configuration whilst trying to find the
problem, however I'll defer any future migrations though until I've got to
the root of it.
--
Ian Cass