Dhcp segfault : any help appreciated!

[Atwood, Robert C]

Dear List Members:

DHCP on our small computational cluster keeps crashing with 

****/var/log/messages ****
Aug  1 16:07:27 mt-hive2 kernel: [670251.004429] dhcpd[15631]: segfault
at 0000000000000000 rip 00002aaaaaed546f rsp 00007fffff944a10 error 4 

After some further time, the nodes start dissappearing from the internal
Net. The people who supplied and configured the cluster don't know why
it crashes
either, so I thought maybe someone here with more DHCP experience could
suggest a course of action to find out.

It is suse-10 based 'ClusterVision OS'  with isc-dhcp-3.0.3 (SUSE
version 3.0.3-2)  I can't see
anything obviously similar on the isc-dhcp mailing list/webpage so far,
but then I don't know what is triggering this. 

I restarted it this time with 
 dhcpd -tf dhcpd.trace -d >& dhcpd.out

in hopes of getting some more information next time it happens, probably
in a couple of days. Of course I don't yet know how to interpret the
trace file to gain useful understanding of what happens.


Does anyone recognize this problem, or can I provide furhter
information? The dhcp is only serving the private network inside the
cluster.

Thanks
Robert