[Linux-cluster] Add a fence device of type SUN ILOM

[Lanthier, Stéphanie]

Message: 6
Date: Fri, 10 Aug 2007 09:32:51 +0200
From: Maciej Bogucki <maciej.bogucki at artegence.com>
Subject: Re: [Linux-cluster] Add a fence device of type SUN ILOM
To: linux clustering <linux-cluster at redhat.com>
Message-ID: <46BC14A3.70303 at artegence.com>
Content-Type: text/plain; charset=UTF-8

> As the GFS file system is not accessible when I'm rebooting one of the 
> three nodes, I'm realizing the importance of fence devices.
It is strange to me, because if You have rc scripts and quorum properly configured, and when You perform reboot of one node Your GFS filesystem should be accesible all time.

Best Regards
Maciej Bogucki

Dear Maciej,

To answer your question, I read this about fence behavior on http://www.centos.org/docs/4/4.5/SAC_Cluster_Suite_Overview/s2-fencing-overview-CSO.html

"When the cluster manager determines that a node has failed, it communicates to other cluster-infrastructure components that the node has failed. The fencing program (either fenced or GULM), when notified of the failure, fences the failed node. Other cluster-infrastructure components determine what actions to take - that is, they perform any recovery that needs to done. For example, DLM and GFS (in a cluster configured with CMAN/DLM), when notified of a node failure, suspend activity until they detect that the fencing program has completed fencing the failed node. Upon confirmation that the failed node is fenced, DLM and GFS perform recovery. DLM releases locks of the failed node; GFS recovers the journal of the failed node."

As I had no fence running, I understood that DLM and GFS were in suspend, waiting to know about the fence completion.

Best regards

__________________

Stephanie Lanthier

Analyste de l'informatique
Universite du Quebec a Montreal
Service de l'informatique et des telecommunications
lanthier.stephanie at uqam.ca
Telephone : 514-987-3000 poste 6106
Bureau : PK-M535