[Linux-cluster] RFC: updating cluster.conf

[Chris St. Pierre]

On Mon, 23 Jun 2008, David Teigland wrote:

> 1. How often do you update cluster.conf?  ("Never" would be valuable
>   feedback.)

Very rarely.  Close to never.

> 2. What changes do you make?  e.g. add nodes, change fencing settings,
>   add or change rgmanager settings.

Since we've just started using RHCS, we're still doing some config
tweaks from time to time.  In the long run, I suspect it will mostly
be to add and remove nodes, and, perhaps more rarely, rgmanager
services.

> 3. How do you currently update cluster.conf?  Cluster online or offline?
>   Manually scp to all nodes?  ccs_tool?  conga?  What do you like and not
>   like about the method you use now?

We put the new cluster.conf in Cfengine and let it copy it out to the
cluster.  When Cfengine detects that it has copied out a new
cluster.conf, it automatically runs the ccs_tool/cman_tool magic to
get the cluster to read the new config.

> 4. How would you like to do updates to cluster.conf in the future?
>   Conga (graphical management interface)?  Command line program that
>   updates /etc/cluster/cluster.conf on all cluster nodes?
>   Manually scp to all nodes?  Other?

I rather like the way we do it now.

> 5. Would you like to use an LDAP server?  All cluster nodes would read
>   cluster.conf info from the server; updates would just be made on
>   the server.

On the one hand, that would be unspeakably awesome.  On the other
hand, that could limit our ability to run an LDAP cluster under RHCS,
although I suppose we could just use the old-skool cluster.conf for
that cluster.

For us, if LDAP is down, we're already hosed beyond hosed, so the idea
of an additional external dependency isn't really that distasteful.
That said, it might be nice to have RHCS cache the config locally to
avoid a temporary LDAP outage from causing a cluster outage, too.

Chris St. Pierre
Unix Systems Administrator
Nebraska Wesleyan University