[ic] Check Out Freezing Up

Zack Johnson interchange-users@interchange.redhat.com
Tue Aug 28 17:02:00 2001


We've been seeing what appears to be the same problem.  In particular,
restarting does not work even in the absence of any /var/run or /var/lock/
entries.  After a few minutes, starting is possible.  We've set a daily cron
job to restart the server, and so far so good.

The whole IC server shuts down, so admin is effected as well.

Single CPU, Stock RH 7.0, stock IC 4.8.1, perl 5.6.0.  Sendmail and DNS are
fine, and we believe, unrelated.

Sometimes we find hammered pid entries, sometimes not.

We're stumped here.

zack
----- Original Message -----
From: "Dan Browning" <db@cyclonehq.dnsalias.net>
To: <interchange-users@interchange.redhat.com>
Sent: Tuesday, August 28, 2001 3:19 PM
Subject: Re: [ic] Check Out Freezing Up


> At 12:53 AM 8/28/2001 +0200, you wrote:
> >Xavier de GELIS wrote:
> > >
> > > Hello,
> > >
> > > On my platform (Celeron 700, 384MB RAM, 20GB, Redhat 7.1, Interchange
> > > 4.8.1), it takes about 25s just to see the check-out page, but it is
not
> > > freezing up. I use the original catalog too.
> >
> >Hmmmm...it is difficult to give you any right answer or solution. I've
> >noticed a long running access of a page too. But I get this behavior
> >only, if there are a few pid files like "pid<PIDNUMBER>" in
> >cataloges/<yourshop>/etc. For example:
> >
> >1.
> >an user call a page and a new process with pid 12345 is starting
> >2.
> >something goes wrong with that process with pid 12345
> >3.
> >closing or shut down the browser, restarting the browser and a new
> >access to the shop once more need more times or "Internal server error"
> >or "document has no data" is showing up suddenly
> >4.
> >after a long time, the pid file 12345 is still in /etc
> >5.
> >restarting the IC server has no effect to get a normal behavior
> >6. removing the pid file in /etc solve it
> >
> >It seems to me, the function housekeeping is not working right or can't
> >remove the wrong pidfiles or has another problem.
> >
> >I don't know about the cause, because it is difficult to reproduce that
> >kind of behavior and to debug it with my less available time. It is
> >possible, that you can access another page quickly. But we need more
> >informations. Can you reproduce the misbehavior? Can you investigate or
> >find a coherence to get such kind of the misbehavior?
> >
> >Joachim
>
> In addition to Joachim's suggestions, what is your platform, distribution
&
> perl version?  Second, what, if any thing, have you modified from the
> foundation catalog?  Third, does your sendmail work correctly (name
> resolution, etc.)?  HTH,
>
> Dan Browning, Cyclone Computer Systems, danpb@mail.com
>
> _______________________________________________
> interchange-users mailing list
> interchange-users@interchange.redhat.com
> http://interchange.redhat.com/mailman/listinfo/interchange-users
>