apache crash after log rotate

[Tomas Hoger <thoger@pobox.sk>]

Hi.

My post is similar to what was discussed in
  http://lists.debian.org/debian-apache/2003/debian-apache-200309/msg00000.html

This has happend to me at least 5 times during a year or so.  Apache died
right after log rotate.  Actually, apache's master process died, some child
processes were still hanging around.  Log notices are very brief ;(

[Tue Dec  2 06:28:20 2003] [notice] SIGUSR1 received.  Doing graceful restart
accept_mutex_on: Identifier removed
accept_mutex_on: Identifier removed
[Tue Dec  2 06:28:20 2003] [notice] Apache configured -- resuming normal operations
[Tue Dec  2 06:28:20 2003] [notice] suEXEC mechanism enabled (wrapper: /usr/lib/apache/suexec)
[Tue Dec  2 06:28:20 2003] [notice] Accept mutex: sysvsem (Default: sysvsem)
[Tue Dec  2 06:28:20 2003] [alert] Child 12842 returned a Fatal error...
Apache is exiting!

Remaining child processes managed to serve few requests, but soon stopped
responding to requests, but they were still alive and bound to ports 80 and
443.

I tried googling for this problem several time, but all I found was mails
describing problem with no relevant answers.  It happens to me too often to
just describe this problem as "lunar ray hitting the power cable during
solar eclipse".  So I ask once again: does anybody have an idea, why this
happens?  Or at least is there any way to get more info about what has
happened?

Apache, I run, is 1.3.26-0woody3 on woody system, with mod_php
4.1.2-6woody3 and mod_ssl 2.8.9-2.1.

Any ideas?

Thanks in advance...

-th.

PS: Please, CC: your answers to my email (thoger@pobox.sk), thanks.