It also relates back to a post I made many moons ago. I don't upgrade any
OS (outside of minor patches). If it is a new OS version, backup and then
clean install. Just too many variables to contend with to trust a type of
upgrade script.
Adam Lang
Systems Engineer
Rutgers Casualty Insurance Company
----- Original Message -----
From: "Steve Wolfe" <steve@iboats.com>
To: <pgsql-general@postgresql.org>
Sent: Tuesday, October 31, 2000 5:03 PM
Subject: Re: [GENERAL] postgres on redhat 7.0
> > I was the original poster. Really my confusion stemmed
> > from the fact that upgrading from RH6.0 to RH7.0, 7.0
> > complained (during boot) that my Postgress verision was outdated
> > I need to upgrade. This threw me off.
>
> Yeah, the init script that they provide checks what's in PG_VERSION, and
> tells you to upgrade. The bad side is that it may have already
overwritten
> your old binaries, etc., making it difficult to do a dump. You'd have to
> reinstall the old one, do the dump, then upgrade PostgreSQL, and then
> reinsert.
>
> Hopefully, their upgrade system was smart enough to not blindly
overwrite
> your old PostgreSQL installation. If it did blindy overwrite it, then
it's
> a very poorly written "upgrade", even Microsoft does better than that in a
> lot of situations. It would give further validation to my refusal to ever
> use RedHat's upgrade procedure.
>
> This isn't to say that RedHat is the devil. Just that like all *nix
> varieties, it has it's own behavioural deficiencies that need to be
> recognized and worked around. If there was a *nix that didn't have
> deficiencies, then all of the other varieties would quickly go away.
>
> > In general I am pretty pissed at RH attitude to system
> > upgrade, if I were working in a Production environment,
> > I would either hire them and not try anything myself,
> > which kinda contradicts the whole Linux philosophy.
>
> Well, it certainly doesn't contradict the RedHat philosophy of "Give
them
> the product for free, then charge for support." ; )
>