Surviving 1 node failure in stretched cluster

[J-Philippe]

Hello, is it possible for some VM that are not anymore protected by HA (storage HA lost sync) to be HA compliant once again even with a node failure.

I mean stretched cluster of 4 nodes (1 & 2 in zone A, 3 & 4 in zone B)

 Node 3 have failed in zone B

VM that were "paired" on the node 3 with one of the node in zone A are ot anymore protected by HA.

If for exemple VM "File" have Primary data on Node 1 and Secondary data on node 3. If Node 1 fails also I will face "DU" for the VM "File". Do I have the possibility to "re dispatch" the data on the node 3 for that VM to the node 4 that is still available ? If yes how can I do it ?

Hope to be clear, thanks.

[DaveOb · Accepted solution]

No.Simplivity maintains 2 copies only  of the data.

A stretched cluster forces the copies to be maintained in distint failure domians.

IE we can survive a catastorphic failure of Zone A or  Zone B

What you are describing if I am correct is a catasrophic failure of a zone combinded with an additional failing node and how to prevent this causing dataloss/unavailability.

There is no way to add a third copy  of the data on either Zone.

A remote backup to another site would be the only way to ensure availability.