[ClusterLabs] Avoid resources to start in parallel

Michael Schwartzkopff ms at sys4.de
Thu Mar 19 06:42:41 EDT 2015


Am Donnerstag, 19. März 2015, 11:20:45 schrieb Jose Manuel Martínez:
> Hello everybody.
> 
> Trying to resolve many problems with Lustre and corosync/pacemaker, I
> would like to know if is there any way to delay the start of a resource
> for a random number of seconds.
> 
> My resources are "filesystems" and if many of then are started at the
> same time at the same node, Lustre randomly fails. At this moment I'm
> trying to work-around it using "Ordering Constraints" in order to
> serialize the resources start order, but this is only a patch. Is there
> any other way to avoid resources to start at the same time?.
> 
> Regards.

Start the reources ordered. Or have a took into the Filesystem resource agent 
and add some random delay.

Beware to fix the start timeout in your configuration.

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, Marc Schiffbauer
Aufsichtsratsvorsitzender: Florian Kirstein
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 230 bytes
Desc: This is a digitally signed message part.
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20150319/89f2c653/attachment-0003.sig>


More information about the Users mailing list