[openstack-dev] [telemetry][ceilometer]when the rabbitmq queue is full of the message. ceilometer-collector is killed.

张盛平 zhangshengping at unitedstack.com
Mon Oct 24 03:11:09 UTC 2016


Hello everyone, 
   I am using Ceilometer (Liberty version). recently, I stopped the ceilometer-collector service for a two or three days, when I start it up again, it is ran a while, then blocked, finally shows Killed (like below).
   
      https://gist.github.com/zhang-shengping/12273667234f440445c5fbe2d71d983b
   
   The event.sample and metering.sample queues are full of messages wait to send. does it cause the ceilometer-collector killed?


[root at server~ ]$ rabbitmqctl list_queues messages name |grep -v ^0
Listing queues ...
133339	event.sample
597079	metering.sample
1	notifications.sample  


     Initially, I though it was rabbitmq problem, I use oslo_messaging notification to connect to the queue and get the samples, I can get the samples quickly. I also debugged a bit, I find it stop at evenlet wait() event, but I dont know why. could you give me some help? thank you very much.


Best regards
Shengping
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20161024/e094e4bb/attachment.html>


More information about the OpenStack-dev mailing list