[Linux-cluster] MySQL + RH Cluster Suite + GFS

Eloy Acosta Toscano eloy.acosta at fon.com
Thu Jun 22 15:46:38 UTC 2006


The problem is that when I power off one node ( I mean disable its power 
supply totally ) to test it, the fencing does not work, on the log I get:

fenced: fencing node "nodename"
fenced: fence "nodename" failed

This will show forerer and the cluster will hang, until the other node join 
the cluster again.
The problem is that, the fenced daemon waits (forever) until the iLO of the 
dead node back to life.

So my cluster is node reboot/reset fault tolerant, but not totally power off 
tolerant.
Do you know what I mean?

Can anybody helps me to workaround the problem?

thak U very much


----- Original Message ----- 
From: "James Parsons" <jparsons at redhat.com>
To: "linux clustering" <linux-cluster at redhat.com>
Sent: Thursday, June 22, 2006 5:17 PM
Subject: Re: [Linux-cluster] MySQL + RH Cluster Suite + GFS


> Eloy Acosta Toscano wrote:
>
>> Hello, I´d like to make you some questions about or new enterprise 
>> database production systems, which should be in production state in about 
>> 2 or 3 days. Yes, I´know this is not too much time to deploy a good and 
>> definitive HA solution, but I´d like to try making the best possible.
>>  We´ve tried a MySQL with RedHat Cluster Suite H/A solution, but we have 
>> very
>> bad results in the H/A tests.
>> We use a iSCSI LUN per node with ext3 FS and a HP iLO as fence device, 
>> and I
>> think that these poor results are related with these two components: ext3
>> formated HD device and a disfunction in the fenced daemon with the HP iLO
>> device.
>
> Tell me about your iLO woes.
> -J
>
>
> --
> Linux-cluster mailing list
> Linux-cluster at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-cluster
> 





More information about the Linux-cluster mailing list