VSAN disk syching back to original failed host

[hongjun]

Assuming one host failed out of 4-host cluster for 60-mins, VSAN will synching datas among 3 remaining hosts and back to compliance mode.

What will happen after the original host coming back online, will VSAN try to do a datastore load balancing and synching some replicas back to the original failed host?

or

it will be happy with fully compliance 3-host layout and leave the newly joined host storage there being idle until a failure happen or new VM storage provisioning? This is assuming default storage policy failure to tolerate=1 and stripe=1.

Thanks

Hongjun

[hongjun]

in my testing, it's the second case. From what I see in Essential VSAN book, it points to second case also.

The problem with this is that assuming no new VM provisioned or no VM storage policy changes, out of 4 host cluster, one host's SSDs and HDDs will be left without any usage, that's 25% performance and capacity hit out of 4-node cluster.

It'll be nice that VSAN has the option to auto scan VSAN nodes and try to balancing replicas among all health nodes.

[hongjun]

any comments?

[williambishop · Accepted solution]

It cleans up after it returns online, and is available capacity for the next fail, create, etc. It is not a failback scenario...unless someone much more experienced comes along to correct me, I will go with Duncan's statements so far on the matter.

[hongjun]

Thanks William.

[williambishop]

It will rebalance if you give it some time, as soon as it hits a watermark, or I think also for a maintenance

[hongjun]

I didn't see it rebalanced and I don't see RVC has any options to manually rebalance hosts storage. It'll be nice to have such an option to manually rebalance the host storage.

[williambishop]

Take a look at maintenance then, you may not be hitting a threshold that triggers a rebalance.