>>> Dejan Muhamedagic <deja...@fastmail.fm> schrieb am 21.09.2012 um 15:16 in
Nachricht <20120921131618.GA25587@walrus.homenet>:
> On Thu, Sep 20, 2012 at 09:35:54AM +0200, Ulrich Windl wrote:
> > (This message was sent to the linux-ha-dev list by mistake (EMail 
> auto-completion fooled me)
> > 
> > > Hi!
> > > 
> > > I have a question on the Xen RA: The docs say:
> > > ---(resource-agents-3.9.3-0.7.1 of SLES 11 SP2)---
> > > node_ip_attribute (string): Node attribute containing target IP address
> > >     In case of a live migration, the system will default to using the IP
> > >     address associated with the hostname via DNS or /etc/hosts.
> > > 
> > >     This parameter allows you to specify a node attribute that will be
> > >     queried instead for the target node, overriding the IP address. This
> > >     allows you to use a dedicated network for live migration traffic to a
> > >     specific node.
> > > 
> > >     Warning: make very sure the IP address does point to the right node. 
> Or
> > >     else the live migration will end up somewhere else, greatly confusing
> > >     the cluster and causing havoc.
> > > ---
> > > 
> > > Now if I have a > 2 node cluster, do I need to specify multiple IP 
> addresses? 
> > > If not which IP address should I specify. I feel the documentation is 
> wrong 
> > > or incomplete.
> 
> The node attribute is per node, so there should be just one IP
> address per node. Something like:
> 
> # crm node attribute mynode set ip myip

Hi!

So I select on unique attribute name for Xen migration, specify that in the Xen 
resource, and then define that attribute per node, using one of the node's own 
IP addresses? I feel that being able to specify an interface (address & 
netmask) (assuming the resolver lists that interface for the node) would be 
much easier.

Regards,
Ulrich


_______________________________________________
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

Reply via email to