[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ha
Subject:    Re: [Linux-HA] clean shutdown procedure?
From:       Arnold Krille <arnold () arnoldarts ! de>
Date:       2013-04-19 19:27:54
Message-ID: 20130419212754.5a883384 () xingu ! arnoldarts ! de
[Download RAW message or body]

[Attachment #2 (multipart/signed)]


On Fri, 19 Apr 2013 08:01:04 -0600 Greg Woods <woods@ucar.edu> wrote:
> Is there a recommended method for taking a cluster out of service
> cleanly?

There are several ways to shutdown the cluster without causing all
resources to fail-over before being shut down again:
 - Send both/all nodes the shutdown command at the same time. For
   example with multiple ssh-connections copying the commands from one
   to the other (cluster-ssh or Konsoles builtin feature).
 - Send both/all nodes the acpi-shutdown at the same time. Either by
   ipmi or by two fingers pressing the power-buttons at the same time.

 - I wouldn't use maintainance-mode. Its a lot of work stoping all the
   services by hand and one of the rules of pacemaker is that you don't
   touch the jobs pacemaker is managing. And if pacemaker can't shut
   down your services, something is wrong...
 - Put the cluster in "stop-all-resources".
 - Use "crm configure" to put all nodes into "standby" at the same
   time. This also leaves the nodes up so you can check the
   non-pacemaker services before shutting each node down individually.
   And you can also select whether on the reboot the nodes should stay
   "standby" or be online again.

I personally use the last one when I have to shut down the whole
cluster at once.

Have fun,

Arnold

["signature.asc" (application/pgp-signature)]

_______________________________________________
Linux-HA mailing list
Linux-HA@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

[prev in list] [next in list] [prev in thread] [next in thread]