On Thu, Aug 24, 2017 at 9:55 PM, Bill James <bill.james@j2.com> wrote:

We have an ovirt master (engine) host in Los Angeles and some remote servers in the UK.
Normally they work fine, but when there is a heavy load on the UK servers the management engine has problems with heartbeat and ends up trying to restart the nodes.

Perhaps the mgmt interface is used for traffic other than mgmt? On small scale it's OK. For bigger scale and workloads, it's best to separate traffic to dedicated NICs.

I saw in this thread that I can change vdsHeartbeatInSeconds (https://www.mail-archive.com/users@ovirt.org/msg41695.html)
but I don't really want to change it globally, just for the nodes in UK.
Also not sure how to get the current setting of that value, only how to change it. How do I tell current value? I heard default is 30 seconds.

To change it:

usr/share/ovirt-engine/dbscripts/engine-psql.sh -c "update vdc_options set option_value = 90 where option_name = 'vdsHeartbeatInSeconds'

ovirt-engine-4.1.0.4-1.el7.centos.noarch

I recommend upgrade, though not specifically due to the above issue.

Or maybe its not best practice to have a cluster that far from the engine?

We have an Engien in Israel managing hosts in Europe and the US.

2017-08-24 11:27:51,921-07 WARN [org.ovirt.engine.core.vdsbroker.VdsManager] (DefaultQuartzScheduler3) [feefbf3f-d0e2-4a64-b008-80838d04f130] Failed to refresh VDS, network error, continuing, vds='ovirt1.evuk.j2noc.com'(d0482635-93fd-4cc3-9c78-523078845f11): VDSGenericException: VDSNetworkException: Heartbeat exceeded

_______________________________________________
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users