[Linux-cluster] Preventing clvmd timeouts
linux at alteeve.com
Thu Jan 26 13:55:45 UTC 2012
On 01/26/2012 08:45 AM, Devrim GÜNDÜZ wrote:
> This is RHEL 6.2, where all packages are up2date.
> For a simple 2-node cluster, clvmd cannot start, because it cannot scan
> clustered LVs -- if one of the nodes are down. It simply hangs in the
> vgscan phase.
> Is that the expected behaviour? If not, what is the recommended way to
> get rid of this?
I am guessing that you do not have fencing configured and/or tested.
When a node is lost, fenced calls DLM which then blocks locking until a
fence call succeeds. Without fencing, this will hang all things using
DLM, including clustered LVM.
E-Mail: digimer at alteeve.com
Papers and Projects: https://alteeve.com
More information about the Linux-cluster