Native SAN. Cluster fail after shutdown of both Nodes

[csauterel]

Hi Starwind Team.

I'm making some testing and I have the following problem.
After a power failure. When the energy comes back, and the two nodes start windows server. The luns are not available because the syncronization fails (I'm using Write Through Cache). If the luns are not available, the Witness lun for the quorum is not available too. And if the lun of the quorum is not available, the cluster will not start, and then the HA-VM also dont start.

What can I do to resolve this??
Dont configure Caché??

PD: If I force full syncronization after the nodes come up, the witness disk comes available, the cluster starts, and the VM starts too.

Best Regards.

[anton (staff)]

First of all - both nodes down is called a DISASTER. Second please use write back cache as write thru affects performance. Third yes force full sync if fast sync is N/A.

Hi Starwind Team.

I'm making some testing and I have the following problem.
After a power failure. When the energy comes back, and the two nodes start windows server. The luns are not available because the syncronization fails (I'm using Write Through Cache). If the luns are not available, the Witness lun for the quorum is not available too. And if the lun of the quorum is not available, the cluster will not start, and then the HA-VM also dont start.

What can I do to resolve this??
Dont configure Caché??

PD: If I force full syncronization after the nodes come up, the witness disk comes available, the cluster starts, and the VM starts too.

Best Regards.