[ClusterLabs] crm_resource --cleanup and cluster-recheck-interval
kgaillot at redhat.com
Thu Jun 23 17:59:31 EDT 2016
On 06/15/2016 05:44 AM, Vladislav Bogdanov wrote:
> It seems that after recent commit which introduces staggered probes
> running 'crm_resource --cleanup' (without --resource) leads to cluster
> to finish recheck too long after cleanup was done. What I see: cluster
> fires probes for the first batch of resources, receives their status,
> writes it to CIB, and then sleep until cluster-recheck-interval lapses.
> After that next batch is reprobed and so on.
> This looks like a regression to me.
> Should I file bug for that or that can be fixed quite fast? IMHO this is
> major enough to be fixed in 1.1.15.
Please file a bug, and attach any relevant logs. Obviously, we weren't
able to handle this in time for 1.1.15, but if it's related to staggered
probes, that was introduced in 1.1.14.
More information about the Users