Greetings,
First let me thank the gentleman who allayed my fears about CentOS 5.1's "yum update" download size. but all the same it turned out to be in megabytes. not to mention the k3b *roasted* couple of DVD's to uselessness and I am still giving explanations to the stupid ancient,retired, former stupid cretin bean counters masquerading as "managers" and "management Accountants" on. a piddly amount wasted.
I found this thread in the GMANE list as I subscribed to this list rather late.
Let me propose a solution:
Let two nodes run the greatest ever CentOS 5.0 on two nodes.
Let two nodes of the cluster run DRBD 0.8 and export block devices in Primary/Primary mode and the GFS be mounted on them.
Let the NFS export be a managed service of the Cluster infrastructure from these nodes exporting the NFS on a cluster managed virtual IP as rw.
let n (as permitted by the infrastructure) nodes import NFS as client on their nodes. and do the job
I am seeing a 3-node cluster exactly that far past fes days and naturally, given the maturity of CentOS 5, do that for, hopefully, ever.
In the ,eantime of the replies hopefully the readers will enjoy the solutions proposed by me
regards,
Rajagopal