[ic] Interchange 4.9.6 process randomly exiting.

[Cameron]

----- Original Message -----
From: "Stefan Hornburg" <racke@linuxia.de>
To: <interchange-users@icdevgroup.org>
Sent: Monday, February 17, 2003 4:04 AM
Subject: Re: [ic] Interchange 4.9.6 process randomly exiting.


> On Mon, 17 Feb 2003 02:24:12 +0800
> "Cameron" <ritontor@icenet.com.au> wrote:
>
> > > > > I've just installed 4.9.6 on a new server, and it's periodically
> > crashing,
> > > > > leaving no trace of why/how/when in the error log(s). Here is the
> > result
> > > > of
> > > > > a strace -p on the process id... It flips along on, looping and
timing
> > out
> > > > > (where it says Timeout) until eventually, at the end, it is dying
:)
> > > > >
> > > > > select(16, [7 8], NULL, NULL, {3, 0})   = 0 (Timeout)
> > > > > time(NULL)                              = 1045380008
> > > > > open("/var/run/interchange",
> > O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY)
> > > > =
> > > > > 2
> > > > > fstat64(2, {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
> > > > > fcntl64(0x2, 0x2, 0x1, 0x8e6e538)       = 0
> > > > > getdents64(0x2, 0x9548b88, 0x1000, 0x4) = 312
> > > > > getdents64(0x2, 0x9548b88, 0x1000, 0x4) = 0
> > > > > close(2)                                = 0
> > > > > select(16, [7 8], NULL, NULL, {3, 0})   = ? ERESTARTNOHAND (To be
> > > > restarted)
> > > > > +++ killed by SIGKILL +++
> > > > >
> > > > >
> > > >
> > > > Further to this, it's happening *EXACTLY* once every hour. i wrote a
12
> > > > second perl script to restart IC when it dies, and yeah... it's
dying
> > almost
> > > > to the second exactly one hour after it's started. This should be a
big
> > clue
> > > > for someone who knows the guts of IC better than i do!
> > >
> > > That looks like Interchange is killed by another program, e.g.
something
> > run
> > > by your provider if this isn't your machine.
> > >
> > > Bye
> > >       Racke
> >
> > well, it is my machine, i got a dedicated server... i've looked in the
> > crontab, nothing too obvious in there. i'll talk to support though, see
if
> > they're doing something deliberate here. i can't imagine what though.
they
> > certainly SHOULDN'T be doing anything...
>
> Did you look in all crontabs ? If you don't find anything there (watch
your
> system logs too !), check on incoming connections at the times where the
> crashes occur.
>
> Bye
>      Racke
>

yeah i poked all the crontabs. it seems unlikely to be a cron job though,
it's happening 1 hour after IC starts, not every hour on the hour. I imagine
that limits.conf would have been a more appropriate place to check - indeed,
i'd already done this before even posting the first time round, but the
entire file is hashed out, so i doubt it's anything in there.