Converting to HA without downtime - is this possible?

[GT80]

Hi there

I have a single StarWind VSAN server with 3 sets of RAID volumes (so 3 sets of MPIO iSCSI targets). 1 volume is currently live while the other 2 have nothing on them. Now I have a newly deployed second server with the same identical hardware. Is it possible to transform the single server setup into a 2 node HA cluster without any downtime by doing the following:

- Set up a HA cluster only for the 2 unused volumes, while the 1st volume remains untouched and live, on the single server.
- Once that is done and heartbeat/sync set up and working, we do a live storage migration from the 1st volume to the 2 newly formed HA volumes
- Then create a new HA cluster for the 1st volume and let the iSCSI clients rediscover/reconnect to the iSCSI target.

And then we should have a full 2 node HA cluster after this? I am quite fearful that if I create a HA cluster for the 2nd/3rd volumes that the 1st volume will somehow stop working or dump its active iSCSI connections.

Please do let me know if this can be done.
Thanks!

[Boris (staff)]

The sequence of actions you described is exactly what you need to do to achieve your goal. The 1st volume will not suffer while the 2nd and the 3rd ones are created.