On Thu, 1 Jul 1999, Philip Hazel wrote:
> On Wed, 30 Jun 1999, Steve Lamb wrote:
>
> > Well, this would explain it. Let me rephrase the question, any reason
> > why Exims from June 10th haven't timed out yet? :/
> >
> > mail 750 0.0 0.0 1576 44 ? S Jun 25 0:01 /usr/sbin/exim -Mc
>
> > root@teleute:~# exiwhat
> > 750 2.05 delivering 10xRLD-0000By-00: waiting for a delivery subprocess to
>
> I have occasionally seen these on Solaris systems, but for *incoming*
> rather than outgoing messages. It is always stuck in read(), and SIGALRM
> doesn't budge it, though all attempts to reproduce the problem have
> failed. It is usually related to a dial-in connection.
>
> Maybe there is some similar problem on output?
>
> > Should I kill these off or does anyone have any bright ideas on anything
> > else I can do to diagnose this?
>
> Try sending SIGALRM first, to see if that helps. If you can attach a
> debugger to find out where they are stuck (i.e. in which system function
> if possible) it might help. My guess would be write(), but maybe there's
> some other problem.
I've seen a few of these on a system that NFS mounts user's mailboxes. If the
fileserver went down, we'd see messages like that completely stuck presumably
jammed in the act of writing to the mailbox - nothing seemed to shift them and
a reboot was required. I didn't have the opportunity to do any debugging, but
if it happens again I'll have a look around.
DHS
-- David Sheryn <D.H.Sheryn@???> Postmaster, Computing Services
-- City University, Northampton Square, London, EC1V 0HB
-- Phone: (+44) 171 477 8000 Direct 0171 477 8196 Fax: 0171 477 8565
--
*** Exim information can be found at
http://www.exim.org/ ***
