I'm setting up a cluster and finding the documentation on shared storage sparse and lacking depth. Where people have managed to set something up, they seem to think having the files (eventually consistent) in more than one location means the job is done.
So I can have some assurance about whatever method I decide on - and in order to confirm that recovery works as expected I want to simulate the sudden loss of a cluster node before I am running any live services from it.
I am expecting to have three different network paths (storage, management and general usage) all known to the cluster and the devices are in a remote-ish datacenter.
Short of pulling out the power cable, is there a way to simulate the loss of a node?
So I can have some assurance about whatever method I decide on - and in order to confirm that recovery works as expected I want to simulate the sudden loss of a cluster node before I am running any live services from it.
I am expecting to have three different network paths (storage, management and general usage) all known to the cluster and the devices are in a remote-ish datacenter.
Short of pulling out the power cable, is there a way to simulate the loss of a node?
