After Upgrade to 2.1, some services of clusters never receive data anymore (stale)

[pdunkler (Paul Dunkler)]

Hi there,

we’ve just upgraded our 2.0 checkmk instance to 2.1.
Everything is working fine except an issue with our varnish cluster checks.

Starting from version 2.1, half of the clustered Varnish-checks are not getting data anymore. They are now marked as stale. Its always the same set of services which don’t get data anymore.
The cluster consists of two hosts.

The check itself doesn’t seem to be broken. For testing i have created two other temporary hosts (same as cluster members). Varnish checks are correctly working on the active cluster host but will stay stale on the cluster.

I did as well try all different possible modes in the new ruleset “Aggregation options for clustered services”. That doesn’t change anything.

Best,
Paul

[davidwayne]

Do you receive data on the participating nodes in the cluster ? You can verify this via “cmk -n or cmk -vII”

[pdunkler (Paul Dunkler)]

Yes, i’ve just checked that both nodes participating in the cluster are correctly receiving the data.
Otherwise, where should the 4 working Varnish services have there data from?

[davidwayne]

@pdunkler

Otherwise, where should the 4 working Varnish services have there data from?

I have seen this behavior in the past. So. just wanted to be sure. Sorry, I have no furhter idea on this apart from looking the cmc.log or web.log with log_level set to debug.

[pdunkler (Paul Dunkler)]

Just tried out the newest available version 2.1. still the same results. Already tried out debug log but can’t find anything…