[Pacemaker] Ressources not moving to node with better connectivity - pingd

Bauer, Stefan (IZLBW Extern) Stefan.Bauer at iz.bwl.de
Mon Dec 9 16:02:23 UTC 2013


Thank you for your time! It works now :)

Stefan

-----Ursprüngliche Nachricht-----
Von: Michael Schwartzkopff [mailto:ms at sys4.de] 
Gesendet: Montag, 9. Dezember 2013 14:26
An: The Pacemaker cluster resource manager
Betreff: Re: [Pacemaker] Ressources not moving to node with better connectivity - pingd

Am Montag, 9. Dezember 2013, 13:06:04 schrieb Bauer, Stefan:
> Why are some resources listed more than once in the output?
> What is the difference between group_color and native_color?
> If a resource has a value of -INFINITY is it because the cluster 
> already decided that this resource should not run on this host or it 
> can not run on this host due to other reasons?
> 
> I'm not quite sure, how a resource stickiness interferes with the 
> internal decicions taken to migrate.

Everything is points. stickiness is points. constraints result in points.

With every event the cluster calculates the matrix from all nodes and all resources. A resource will run on that node it can collect most points.

Beware of impicit constraints that give points. I.e. a colocation

col col_A_with_B inf: A B

will result in -inf points for the resource B on all nodes where A is not running.

Look at the output of crm_simulate -s -L, write down the matrix and understand it.

--
Mit freundlichen Grüßen,

Michael Schwartzkopff

--
[*] sys4 AG

http://sys4.de, +49 (89) 30 90 46 64, +49 (162) 165 0044 Franziskanerstraße 15, 81669 München

Sitz der Gesellschaft: München, Amtsgericht München: HRB 199263
Vorstand: Patrick Ben Koetter, Axel von der Ohe, Marc Schiffbauer
Aufsichtsratsvorsitzender: Florian Kirstein




More information about the Pacemaker mailing list