Monitoring hosts

Community Support
1

Hi,

How I can set a minimum polling interval to monitoring hosts?

e.g.
I have 3 servers with shared storage.

Host hook enabled:

HOST_HOOK = [
    NAME      = "error",
    ON        = "ERROR",
    COMMAND   = "ft/host_error.rb",
    ARGUMENTS = "$ID -m -p 5",
    REMOTE    = "no" ]

I restart one of the servers, but VMs doesnt migrate, because HOST isnt in ERROR status.
But, I need minimal - 20 seconds to check HOST status.

I found this topic, but can`t understand how it can help me.

2

This is set in oned.conf:

http://docs.opennebula.org/5.2/deployment/open_cloud_host_setup/monitoring.html

You are interested in the -i argument for the collectd:

-i: Time in seconds of the monitorization push cycle. This parameter must be smaller than MONITORING_INTERVAL (see below), otherwise push monitorization will not be effective.

3

Thank you for your reply.

Its work, but I have a new problem. Im imitating the host problem (down interface), host in ERROR state, but HOST_HOOK doesn`t execute migrate VMs to another host.

VMs are still running on ERROR host, doesn`t ping, but in RUNNING status.

Скриншот 2017-02-03 03.22.43.png939×173 12.4 KB

Скриншот 2017-02-03 03.22.49.png965×172 16.7 KB

crmsh migrate services to another host.

4

Is there any idea?

5
6

Yes. Now it works.
Big thanks.

Now i try to minimise downtime)