Hi Jochen, thanks a lot for your answer.
I ran snmpd via strace with the same parameters that /etc/init.d/ uses. A local snmpwalk works as usual, but the remote one was stuck from the beginning. This would support your conjecture that it's kernel related. The strace output of the unanswered remote snmpwalk can be found here: http://www.phys.ethz.ch/~daduke/snmpd_no_response.log I hope you'll be able to read something out of it, because I'm not. One additional piece of information that might be of interest: as already mentioned, this is a Xen dom0. Two weeks ago the lenny Xen kernel was not stable, hence I used etch's 2.6.18. The latest 2.6.26-10 seems to run better and I switched back. The snmp problem is present with both kernels on the lenny system. However, servers running etch and 2.6.18 with the same snmpd config respond to snmpwalks just fine. Please let me know if I can be of any further assistance. thanks, -Christian On Tue, Dec 02, 2008 at 07:00:44PM +0100, Jochen Friedrich wrote: > Hi Christian, > > > after a random amount of time (hours to a few days), snmpd stops responding > > to remote requests. Local requests continue to work. A restart of snmpd does > > not solve the problem, only a reboot does. > > Could yo run snmpd -f via strace in such a case? If a restart of snmpd > doesn't help > it looks like the problem could be caused by some kernel settings the snmpd > daemon > tries to access (maybe some unexpected proc variable contents or similar). > > Thanks, > Jochen -- Dr. Christian Herzog e-mail: [EMAIL PROTECTED] IT Systems Specialist voice: +41 44 633 3950 Department of Physics office: HPT D 17 Swiss Federal Institute of Technology 8093 Zurich, Switzerland -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
