[ClusterLabs] resource will not move

Tomas Jelinek tojeline at redhat.com
Mon Feb 2 09:34:17 UTC 2026


Dne 01. 02. 26 v 8:44 lejeczek via Users napsal(a):
> hi guys.
>
> I had a "partial" cluster electricity power hard crash - now a resource
> will not move:
>
> -> $ pcs constraint ref MONERO
> Resource: MONERO
>    colocation-MONERO-GATEWAY-NM-link-INFINITY
>    colocation_set_ceMO
>
> that 'colocation_set_ceMO' is quite weird is - is it not? perhaps for
> @devel to poke at it if they think it is weird.

Hi lejeczek,

What do you think is weird about the constraint? It looks like a normal 
colocation set constraint to me.

> ...
>    resource 'MONERO' with resource 'GATEWAY-NM-link' (id: colocation-
> MONERO-GATEWAY-NM-link-INFINITY)
>      score=666
> ...
>    Set Constraint: colocation_set_ceMO
>      score=-1000
>      Resource Set: colocation_set_ceMO_set
>        Resources: 'MONERO', 'compute'
> ...
>
> _colocation_set_ceMO_ was cluster-created, as opposed to by-human - I> removed it and MONERO got moved as expected - why was such constraint
> cluster-created?

Why do you think the constraint was cluster-created?

I'm not aware of a way in which pacemaker or pcs would create such a 
constraint on their own. Are you sure there was no  input from a human 
or a script running pcs / cibadmin to create such a constraint?

Regards,
Tomas

> those two are not in any relationship within PCS at all - MONERO is
> systemd service, 'compute' is a VM - could such crash make cluster so
> subtly & specifically "misbehave"...
>
> thanks, L.
>
> _______________________________________________
> Manage your subscription:
> https://lists.clusterlabs.org/mailman/listinfo/users
>
> ClusterLabs home: https://www.clusterlabs.org/



More information about the Users mailing list