[Linux-cluster] Cluster IP alias doesn't work

[Lon Hohberger]

On Thu, 2005-07-21 at 14:54 -0400, haydar Ali wrote:

> I created a NFS share on /u04 (mount on /dev/sdd4) using the following 
> command cluadmin on RAC1
> I launched on 2 nodes Rac1 and RAC2 the following command:
> Service cluster start
> I checked the result on the 2 nodes:
> clustat

Ok, that config looks good.

> And I launched on RAC1 and other servers the following command:
> mount t nfs 192.168.253.20:/u04 /u04
> 
> And all are OK, I can list the /u04 content from any server (if I mount it).

Good.

> But my only problem is:
> 
> When I want to try a test, I stop the network service on RAC2:
> ifconfig eth0 down
> Then when I try from another server as RAC1 to list the /u04 content, it 
> doesnt work and doesnt respond, and when I ping the IP alias 
> 192.168.253.20 it doesnt respond also.
> 
> Have you any idea to fix this problem?

That's expected behavior...

The IP addresses are all bound to eth0. (eth0:0, eth0:1, etc...)  If you
bring the entire interface down, all service IPs will also go down.

RHEL 2.1's clumanager does not perform link monitoring for services.
You'll have to use the "mon" utility to get it working if you want it.

-- Lon