Problem with dnsmasq getting killed.

asked 2014-06-26 10:08:48 -0600

Akshat gravatar image


We have a openstack environment with 1 controller , 1 network node and many computes, there are multiple such environments some running grizzly and some running havana.

We are facing a common problem everywhere, the dnsmasq process dies on it own and we have to restart the dhcp agent or add the network back to the dhcp agent in case it dies.

We have a lease period of 4 hours and running a load of approx. 800+ VMs.

Has someone faced a similar problem if not any suggestions/pointers will really be helpful.

edit retag flag offensive close merge delete


Can you pull the log data from DHCP and Neutron-Server. Do you see any errors when you match up timestamps with when the process fails? Are you using a single subnet for all 800 VMs?

mpetason gravatar imagempetason ( 2014-06-26 12:07:26 -0600 )edit

Unfortunately there is no error in the logs....yes, we are using a single subnet which has multiple cidrs.

Akshat gravatar imageAkshat ( 2014-06-26 12:23:48 -0600 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2014-06-26 12:28:23 -0600

mpetason gravatar image

Since you are restarting it regularly, set the log options for dhcp to Debug. When the process dies does it die but has a PID associated with it still? You may be able to see what it's getting stuck on looking in /proc/pid/. Is there anything in dmesg/syslog?

edit flag offensive delete link more


You can both edit your question and answer while you go on debugging the issue: this makes this site a lot more useful than to parse brief comments.

smaffulli gravatar imagesmaffulli ( 2014-06-27 16:36:42 -0600 )edit

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2014-06-26 10:08:48 -0600

Seen: 375 times

Last updated: Jun 26 '14