Thread: 8.4 process killed- what sort of recovery options do I have?

For whatever reason- investigations still pending- the postgres process froze up and I had to do a killall b/c pg_ctlcluster would not revive the thing. Everything seems to be in working order, but I’d like to know if there’s any kind of consistency check I should run in light of events. Also, what’s the best way to track down the reason for the crash? Just get into /var/log/postgresql and poke around?

Thanks.

--

Wells Oliver

Developer, Baseball Systems

San Diego Padres | 100 Park Boulevard | San Diego CA 92101

619-795-5359

Wells Oliver <woliver@padres.com> writes:
> For whatever reason- investigations still pending- the postgres process
> froze up and I had to do a killall b/c pg_ctlcluster would not revive
> the thing. Everything seems to be in working order, but I'd like to know
> if there's any kind of consistency check I should run in light of
> events. Also, what's the best way to track down the reason for the
> crash? Just get into /var/log/postgresql and poke around?

I don't think you need to worry about your data.  Crash recovery is a
well-tested aspect of Postgres ;-).  As for figuring out what happened,
did you force core dumps?  Stack traces from the dumps would be most
useful.

            regards, tom lane