Hi,
I`ve heard that it is much safer, to setup separate DRBD resource for each single VM. So each VM has its own LV with DRBD. Is this really safer setup? It is much more complicated and more work/configuring has to be done :( But when some bad things happen to cluster (2 node cluster, HA) would this kind of setup lead to _higher_probability_ that in case of split-brain, it will hit just some of DRBD instances? So at least some of VMs would survive without any problem.
If there would be just one DRBD instance with a lot of LV volumes with VMs on them, split-brain would affect all of them.
Thank you
Pep.
I`ve heard that it is much safer, to setup separate DRBD resource for each single VM. So each VM has its own LV with DRBD. Is this really safer setup? It is much more complicated and more work/configuring has to be done :( But when some bad things happen to cluster (2 node cluster, HA) would this kind of setup lead to _higher_probability_ that in case of split-brain, it will hit just some of DRBD instances? So at least some of VMs would survive without any problem.
If there would be just one DRBD instance with a lot of LV volumes with VMs on them, split-brain would affect all of them.
Thank you
Pep.