On Wed, 3 May 2000, Vince Vielhaber wrote:
> On Tue, 2 May 2000, The Hermit Hacker wrote:
>
> >
> > As Don asks, what happened with the v7.0 trials you were doing? Corrupted
> > indices, I've seen occasionally in older versions, but I can't recall ever
> > seeing corrupt system tables ...
> >
> > I don't have a GUI browser right, so searching the archives is kinda tough
> > for me :( Can you refresh my memory for me? There has to be something
> > logical to this, as to what the cause for the corruption is :(
> >
> > >From Don's comment, I take it you are using FreeBSD? Version? Stability
> > of the machine? Never crashes?
> >
> > Version of PostgreSQL? Compile/configure options? Do you have any core
> > files in your data/base/* hierarchy that would be the result of a backend
> > crashing?
> >
> > I know you are looking at alternatives, but I'm terrible at letting go of
> > problems :(
>
> His description of table corruption and the system running slower and
> slower sounds like a disk going bad. I've seen it hundreds of times
> on news machines. Constant retries while trying to write to the disk
> will give slowdowns. Having data on a spot of the disk that's unreliable
> will certainly cause data integrity problems.
That was one thing I was thinking ... the other was the possibility that
he's mount'd async and his machine is rebooting ... *or* he has memory
problems causing the shared memory to corrupt, dump the postmaster process
which is corrupting his tables ...
Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
