Cluster reconfiguration

[Jamie]

Hi All,

I'm currently working with a large cluster ( > 50 nodes ) so periodically one or more machines may crash or become otherwise unreachable from the head node. I'm looking to remove the offending node from the cluster in this situation such that the head node will stop waiting for a response from it.

I'm just curious if anyone else has dealt with this type of reconfiguration. Is it as simple as "NetworkManager::removeNode(badNode)" or is there other reconfiguration required? 

Thanks,

Jamie