[Linux-cluster] Halt nodes in cluster with cable disconnect

emmanuel segura emi2fast at gmail.com
Sat Jan 28 17:55:45 UTC 2012


It's no true the node without the delay always win

this is from a comment redhat support about this
Red Hat Support  says:

This kbase is a little confusing....The node *without* the fence delay will
always "win" the fence race

It would probably be better to state it this way....

The node without fencedevice "delay" assigned will get fenced faster. So
for the one that you want to prevent getting fenced you'll want that
node's fencedevice to have the delay set.


2012/1/27 Digimer <linux at alteeve.com>

> On 01/27/2012 04:43 PM, Miguel Angel Guerrero wrote:
> > Digimer
> > the echo c > /proc/sysrq-trigger; command works fine the node with
> > this command reboot thank to the fence-peer :) in a scenary without
> > "fencing race", how the cluster take a decisition about what node
> > reboot in the cable disconnection test?
> > One question you think drbd works better in a pacemaker or cman
> environment?
> >
> > Emmanuel
> > Your english is good, i preffer talk in spanish :P sorry for my bad
> > english ever i learn so much thanks to this thread
> >
> > You never say me nothing about my delay fence line
> > <fence_daemon clean_start="0" post_fail_delay="10" post_join_delay="30"/>
> >
> > Digimer And Emmanuel Thanks a lot for your help and patience
> In a fence race, the node with the sleep will always lose.
> DRBD works equally fine with Pacemaker and RHCS.
> --
> Digimer
> E-Mail:              digimer at alteeve.com
> Papers and Projects: https://alteeve.com
> --
> Linux-cluster mailing list
> Linux-cluster at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-cluster

esta es mi vida e me la vivo hasta que dios quiera
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20120128/6e3b7f7f/attachment.htm>

More information about the Linux-cluster mailing list