Monitoring health of management servers

[bradh352]

What's the best way to monitor the health of management servers? I'm running 4.21.0 and there times when one of the management servers gets 'hung', I can tell this because if I restart a host agent it won't reconnect until the afflicted management server daemon is restarted.

I'm honestly not even sure which management server is the cause or how to determine that. I'm still in the early stages of cloudstack and the logs are pretty verbose and I'm not sure what to possible grep for.

I also wonder if maybe the 4.21 release wasn't a good option right now, I guess its not an LTS release and I'm not sure how many people actually test and run those. Is this possibly a known issue in 4.21 where the management servers need to be restarted at some frequency? Sort of thinking of adding a cron job to each management server to restart in the middle of the night.

[rajujith]

@bradh352 If you are using KVM hypervisor make sure to enable the agent LB so that the agents connect to next management server: https://www.shapeblue.com/software-based-agent-lb-for-cloudstack/

Using 4.22 or 4.20.2 would be ideal.
For general monitoring you can refer: https://www.shapeblue.com/monitoring-apache-cloudstack-hypervisors/

[bradh352]

@bradh352 If you are using KVM hypervisor make sure to enable the agent LB so that the agents connect to next management server: https://www.shapeblue.com/software-based-agent-lb-for-cloudstack/

Using 4.22 or 4.20.2 would be ideal. For general monitoring you can refer: https://www.shapeblue.com/monitoring-apache-cloudstack-hypervisors/

Yes, I'm using KVM and I have roundrobin enabled.

Thank, I'll look at that link you provided. I doubt I can downgrade, hopefully 4.22 will fix it once its out.