Query on vsphere 6 HA Behaviour in & after network partition

[PCHANDU]

Dear team...

Good day ....

Can anyone explain the below senario please ....

A 8 node HA enabled cluster has fallen under network partitioned situation. It splitted as 4 & 4. group A and group B.

As per HA funda group A will have one master and due to re election process one more master will be elected in group B.

Until issue resolves each master will have their own protection list of vms and respective lock mechanism.

Once the issue resolves. What will happens exactly. How one of the existing master will be degraded as slave.

What will happens to the VM protection list managed by the both masters. How group B vms locks will be opening.

Please explain in detail .....

Thanks in advance ....

[daphnissov]

Please see this document for an explanation: https://ha.yellow-bricks.com/fundamental_concepts.html

[depping]

I think I blogged about it, or wrote about it in the book somewhere, but can't find it: When the partition is lifted 1 master will give up it's responsibilities, it does this based on the same algorithm used to determine who is a master in the first place (who has access to all / the most datastores, which one has the highest number). One wins, the other one then becomes slave again. The master already automatically periodically scans the datastores for ownership. As soon as the other master drops responsibilities the "new" master will claim ownership.