[Linux-cluster] updating running cluster.conf

[David Teigland]

On Thu, Oct 28, 2004 at 01:50:12PM -0700, chekov at ucla.edu wrote:
> I've been running a two-node cluster for a month or so and finally got 
> myself a couple of APC master switches to perform automatic fencing.
> I want to update my cluster.conf file to reflect the new fencing method but 
> I'd like to do it without an interruption in service.
> When I join a node the cluster simply copies over the old cluster.conf file 
> to the new nodes /etc/cluster directory.  So simply joining a new node with 
> a new config file doesn't do it.
> I tried using "cman_tool version -r" in an attempt to update the conf 
> version.  However now it just rejects node addition in general (and does 
> not update the running cluster.conf file)...
> 
> in short, I don't know how to introduce the new changes without taking down 
> all nodes.  any ideas?

1. have both nodes running as cluster members using the original
   cluster.conf

2. on one node, update /etc/cluster/cluster.conf, incrementing
   config_version

3. on this same node run "killall -HUP ccsd"

4. verify that the new cluster.conf exists on both nodes

5. on this same node run "cman_tool -r <new config_version>"

6. check /proc/cluster/status to verify the new config version

I just tried this and it works with the current code in cvs.  If it fails
for you you may need to update (which probably requires upgrading both
nodes together making it a moot point.)

-- 
Dave Teigland  <teigland at redhat.com>