Re: [GENERAL] Backend cache invalidation initialization failed... HUH?

At 23:15 +0200 on 6/12/98, The Hermit Hacker wrote:

> > Don't be very sure... So far, I have not managed to successfully install
> > 6.4 on my Solaris 2.5.1 (has very wierd regression results), and I saw
> > another complaint to the same effect (can't create database) from another
> > Solaris 2.5.1 user, who also claimed to have had problems with PG 6.3 (I
> > never got around to installing 6.3 so I wouldn't know). Now, this is the
> > third complaint about Solaris 2.5.1 which I see, and I have a bad feeling
> > some factor was introduced with 6.3, which broke Solaris 2.5.1
> > compatibility.
> >
> > Which would mean that I'll have to say goodbye to Postgres, especially
> > since the mailing lists (I posted mine to this [GENERAL] list as well as
> > the ADMIN list and got no response of any kind).
>
> I don't remember seeing it...can you resend or point to it in the mailing
> list archives?
>
> I did all my testing on Solaris 2.6 (x86 and sparc)...didn't figure there
> would be that much of a difference :(

OK, first, I apologise if my message was a bit on the sour side. But our
Sysadmin smirks all over when he gets proof that "It can be done with
Oracle and not with Postgres", and adequete support is a strong one for his
side...

Anyway, I have posted my original problem, that with the strange regression
results, on November 17th, to the ADMIN mailing list.

   http://www.PostgreSQL.ORG/mhonarc/pgsql-admin/1998-11/msg00066.html

Marc, the reason I didn't post it to PORTS to begin with is that I thought
PORTS was dead. After all, it doesn't appear in the mailing lists menu in
the PostgreSQL site...

I should say that I don't know whether my problem has connection with the
two other Solaris 2.5.1 complaints. I just know I need to upgrade, and I
need the new version to pass the regression tests reasonably (concessions
being made for floating point, random and certain time conversions).

It's a dual-processor sparc machine. We have installed flex. Having parsing
errors in the regression, we have installed bison as well and built from
scratch. The situation somewhat improved, but not quite. Anyways, my
sysadmin insists on using the SUNWspro C compiler, having some deeply
ingrained distrust of gcc. That's my environment description, if it's any
help.

Thank you,
Herouth

- --
Herouth Maoz, Internet developer.
Open University of Israel - Telem project
http://telem.openu.ac.il/~herutma