Thread: Sourceforge PG crash

Sourceforge PG crash

From
"August Zajonc"
Date:
http://sourceforge.net/forum/forum.php?forum_id=80610

Be interesting to get some more details on this. Sourceforge is one of more
visable PG sites out there...

AZ




Re: Sourceforge PG crash

From
Justin Clift
Date:
Hi,

From what I know, Sourceforge was running an early PostgreSQL 7.1 beta
or release candidate.  Does anyone know if the crash was due to a known
bug, or something which is unknown and needs to be fixed?

Regards and best wishes,

Justin Clift

August Zajonc wrote:
>
> http://sourceforge.net/forum/forum.php?forum_id=80610
>
> Be interesting to get some more details on this. Sourceforge is one of more
> visable PG sites out there...
>
> AZ
>
> ---------------------------(end of broadcast)---------------------------
> TIP 5: Have you checked our extensive FAQ?
>
> http://www.postgresql.org/users-lounge/docs/faq.html

--
"My grandfather once told me that there are two kinds of people: those
who work and those who take the credit. He told me to try to be in the
first group; there was less competition there."
     - Indira Gandhi

Re: Sourceforge PG crash

From
The Hermit Hacker
Date:
They were running a Beta2, or something like that, database and were even
warned against it by several of us ... I don't know if they ever upgraded
to the full release, but, considering the size of Sourceforge, I'm
doubting it, since its only been a week ... then again, maybe they used
the crash as an opportunity?

On Fri, 20 Apr 2001, August Zajonc wrote:

> http://sourceforge.net/forum/forum.php?forum_id=80610
>
> Be interesting to get some more details on this. Sourceforge is one of more
> visable PG sites out there...
>
> AZ
>
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 5: Have you checked our extensive FAQ?
>
> http://www.postgresql.org/users-lounge/docs/faq.html
>

Marc G. Fournier                   ICQ#7615664               IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: scrappy@hub.org           secondary: scrappy@{freebsd|postgresql}.org



Re: Sourceforge PG crash

From
"August Zajonc"
Date:
Ahhh...

The strong record of stability tempts even the most cautious user into an
early upgrade.
PostgreSQL clearly needs to introduce more instabilities and bugs in the
beta's. I'll be submitting some patches shorty.

August

----- Original Message -----
From: "The Hermit Hacker" <scrappy@hub.org>
To: "August Zajonc" <junk-postgre@aontic.com>
Cc: <bigdisk@users.sourceforge.net>; <pgsql-general@postgresql.org>
Sent: Sunday, April 22, 2001 1:14 AM
Subject: Re: [GENERAL] Sourceforge PG crash


>
> They were running a Beta2, or something like that, database and were even
> warned against it by several of us ... I don't know if they ever upgraded
> to the full release, but, considering the size of Sourceforge, I'm
> doubting it, since its only been a week ... then again, maybe they used
> the crash as an opportunity?
>
> On Fri, 20 Apr 2001, August Zajonc wrote:
>
> > http://sourceforge.net/forum/forum.php?forum_id=80610
> >
> > Be interesting to get some more details on this. Sourceforge is one of
more
> > visable PG sites out there...
> >
> > AZ
> >
> >
> >
> >
> > ---------------------------(end of broadcast)---------------------------
> > TIP 5: Have you checked our extensive FAQ?
> >
> > http://www.postgresql.org/users-lounge/docs/faq.html
> >
>
> Marc G. Fournier                   ICQ#7615664               IRC Nick:
Scrappy
> Systems Administrator @ hub.org
> primary: scrappy@hub.org           secondary:
scrappy@{freebsd|postgresql}.org
>
>


Re: Sourceforge PG crash

From
Tim Perdue
Date:
On Sun, Apr 22, 2001 at 02:14:31AM -0300, The Hermit Hacker wrote:
>
> They were running a Beta2, or something like that, database and were even
> warned against it by several of us ...

That's probably overstating things a bit, but running betas and CVS snapshots
is asking for trouble and we know what we're doing and why we had to do it.

> to the full release, but, considering the size of Sourceforge, I'm
> doubting it, since its only been a week ... then again, maybe they used
> the crash as an opportunity?
>
> On Fri, 20 Apr 2001, August Zajonc wrote:
>
> > http://sourceforge.net/forum/forum.php?forum_id=80610
> >
> > Be interesting to get some more details on this. Sourceforge is one of more
> > visable PG sites out there...

Actually it was beta6 that wigged out and wouldn't start up again.

This is from the PG server log:

The Data Base System is starting up
DEBUG:  ReadRecord: invalid magic number 0000 in logfile 49 seg 121 off
9666560
DEBUG:  redo done at (49, 2039709660)
FATAL 2:  XLogWrite: write request is past end of log
/usr/local/pgsql/bin/postmaster: Startup proc 577 exited with status 512 -
abort

The biggest mistake we've ever made was upgrading from our very very happy
November 17 pre-beta CVS snapshot to B6 - it was just a nitemare that
hosed literally several times a day until it finally conked out the other day.

7.1 released seems very happy so far on our linux 2.4 kernel. I think
something very subtle was wrong with b6, because it gave us serious problems.

Tim

--
Founder - PHPBuilder.com / Geocrawler.com
Lead Developer - SourceForge
VA Linux Systems

Re: Sourceforge PG crash

From
Tom Lane
Date:
Tim Perdue <tim@sourceforge.net> writes:
> Actually it was beta6 that wigged out and wouldn't start up again.

> This is from the PG server log:

> The Data Base System is starting up
> DEBUG:  ReadRecord: invalid magic number 0000 in logfile 49 seg 121 off
> 9666560
> DEBUG:  redo done at (49, 2039709660)
> FATAL 2:  XLogWrite: write request is past end of log
> /usr/local/pgsql/bin/postmaster: Startup proc 577 exited with status 512 -
> abort

This looks like the problem Vadim found and fixed between RC2 and RC3
(restart after crash would fail if next xlog write point was just at a
page boundary).

> The biggest mistake we've ever made was upgrading from our very very
> happy November 17 pre-beta CVS snapshot to B6 - it was just a nitemare
> that hosed literally several times a day until it finally conked out
> the other day.

AFAIR we didn't hear a word from you about problems --- you really
should have let us know about these crashes...

> 7.1 released seems very happy so far on our linux 2.4 kernel.

Glad to hear it.  Please do report any further problems you see.

            regards, tom lane