Mailing List Archive

On 2008-01-26T21:29:39, Andreas Mock <Andreas.Mock@web.de> wrote:

> the crm.dtd says about the "ordered"-attribute:
> -------------------------8<--------------------------
> ordered
> * Start (or stop) each clone only after the
> operation on the previous clone completed.
> -------------------------8<--------------------------
>
> Does anyone know if this is also true for monitor action?

No, monitors are independent from each other. They also don't receive
the notify attributes (nor are pre/post notifies sent for them) - that
would be too much effort to synchronize and replicate.


Regards,
Lars

--
Teamlead Kernel, SuSE Labs, Research and Development
SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde


_______________________________________________
Pacemaker mailing list
Pacemaker@clusterlabs.org
http://list.clusterlabs.org/mailman/listinfo/pacemaker

> -----Ursprüngliche Nachricht-----
> Von: linux-ha-bounces@lists.linux-ha.org
> [mailto:linux-ha-bounces@lists.linux-ha.org] Im Auftrag von
> Lars Marowsky-Bree
> Gesendet: Samstag, 26. Januar 2008 21:41
> An: The Pacemaker cluster resource manager; 'General Linux-HA
> mailing list'
> Betreff: [Linux-HA] Re: [Pacemaker] Ordered attribute of
> clone resources

> No, monitors are independent from each other. They also don't receive
> the notify attributes (nor are pre/post notifies sent for them) - that
> would be too much effort to synchronize and replicate.
>

Hi Lars,

this can have an impact on monitoring stonith resource agents if you
want to use clones. There are many stonith devices which do allow only
extactly one connection at a time. The monitor action is often implemented
to connect to the stonith device to see if it's alive.
Now if you use clones, and the stonith RA is run on more then one node
it could happen that the monitor action is triggered at the same time
resulting in one monitor action to fail because two instances try to connect
to the same stonith device.

Am I wrong with that conclusion?

Best regards
Andreas Mock


_______________________________________________
Pacemaker mailing list
Pacemaker@clusterlabs.org
http://list.clusterlabs.org/mailman/listinfo/pacemaker

On 2008-01-26T22:32:21, Andreas Mock <Andreas.Mock@web.de> wrote:

> this can have an impact on monitoring stonith resource agents if you
> want to use clones. There are many stonith devices which do allow only
> extactly one connection at a time. The monitor action is often implemented
> to connect to the stonith device to see if it's alive.
> Now if you use clones, and the stonith RA is run on more then one node
> it could happen that the monitor action is triggered at the same time
> resulting in one monitor action to fail because two instances try to connect
> to the same stonith device.
>
> Am I wrong with that conclusion?

No, you are not.

Such a device should either not be handled by a clone.

Or instead by an m/s resource - where only the master performs the
monitor; but this gives you maximum complexity while not providing any
benefit over handling it as a primitive resource - fancy, though ;-)


Regards,
Lars

--
Teamlead Kernel, SuSE Labs, Research and Development
SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde


_______________________________________________
Pacemaker mailing list
Pacemaker@clusterlabs.org
http://list.clusterlabs.org/mailman/listinfo/pacemaker