Re: postgres crash SOS

On Thu, 2010-06-17 at 19:27 +0000, Dann Corbit wrote:
> > -----Original Message-----
> > From: Tom Lane [mailto:tgl(at)sss(dot)pgh(dot)pa(dot)us]
> > Sent: Thursday, June 17, 2010 12:20 PM
> > To: Dann Corbit
> > Cc: Felde Norbert; pgsql-general(at)postgresql(dot)org; Scott Marlowe
> > Subject: Re: [GENERAL] postgres crash SOS
> >
> > Dann Corbit <DCorbit(at)connx(dot)com> writes:
> > >> (Personally I'd never run a database I cared about on Windows.)
> >
> > > Somehow, I doubt that Windows is to blame. For instance, Oracle and
> > SQL*Server seem to run fine on Windows without this sort of problem.
> >
> > Really? Are you front-line support for either, so that you can say
> > that they "never" have this sort of problem on the basis of any actual
> > evidence? Our failure rate on Windows is probably pretty low too,
> > it's just that we only hear about non-working cases here.
>
> I have worked as a DBA for large farms of database systems.

I think this argument as a whole represents so many "What ifs" it is not
even worth having.

Fact: Windows Server is a decent OS for databases
Caveat: If properly managed
Wish: A decent shell to work in (although powershell is interesting)

Big fark of a caveat there. Of course the same caveat applies to Linux
or FreeBSD too.

>
> > (Perhaps more to the point, if they don't have problems, it's likely
> > because they tell their customers how to configure Windows boxes safely
> > before the fact. And people who are spending the money for an Oracle
> > license will heed that advice.)
> >
>
> Care was taken in selection of hardware and configuration. Is this different for any database system?

Yes :) but does it matter?

Sincerely,

Joshua D. Drake

--
PostgreSQL.org Major Contributor
Command Prompt, Inc: http://www.commandprompt.com/ - 509.416.6579
Consulting, Training, Support, Custom Development, Engineering