From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; .NET CLR 1.0.3705) Description of problem: the snmpd service on more then 1 of our redhat 8.0 servers has been randomly dying. The config (snmpd.conf) is the same and on many redhat 8.0 machines that work fine. It's been happening on more and more systems as time goes by. Version-Release number of selected component (if applicable): net-snmp-5.0.6-8.80.2 How reproducible: Sometimes Steps to Reproduce: 1. service snmpd start 2. wait.... 3. snmpd will die Expected Results: serivce dies Additional info: This is our workaround for it, but this is a hack and shouldn't be needed: /bin/ps auxww | /bin/grep snmpd | /bin/grep -v check | /bin/grep -v grep > /dev/null || /sbin/service snmpd start
Here's a few suggestions: 1. What specific queries do you interrogate snmpd with? 2. Does it crash randomly or is there a pattern (every 5-6 hrs, 3-4 days?) 3. Do you have IPv6 enabled? (try "lsmod | grep ipv6" or see if /proc/sys/net/ipv6 exists) 4. Read bug #85071 and see if it applies to you :) 5. If IPv6 isn't enabled, try reproducing by querying the same things over and over manually (ie. I'm assuming you run somekind of network management software against your snmpd.) last but not least ... 5. strace the snmpd process until it dies or attach gdb and wait until it dies, do "bt" and report it back..
Putting in needinfo. :-) Read ya, Phil
No input in half a year, closing this bug as worksforme (we have quite a few machines with net-snmp up and running here and i haven't seen any failures as of yet). Read ya, Phil