Two node vSAN with witness, when putting 1 node to maintenance mode with full migration gives 'General vSAN error'

[ChrisFD2]

Hi all,

I have two all flash vSAN nodes with a separate witness.

All hosts running 6.7 build 9484548.

vCSA is 6.7 build 9451876.

If I try and put one of the two into maintenance mode with full data migration I get a 'general VSAN error' and the task fails. If I choose Ensure accessibility it works fine.

Where can I find logs to show me what's going on?

This is a home lab, not production - so no support.

vSAN policy is standard cluster 1 node failure. Should I change it to a stretched cluster policy? All nodes are on the same subnet, the two hosts have two 10 Gb links directly connected and witness traffic is tagged on the network facing VMkernel interface as it should be. Or am I missing something really obvious?

Thanks!

[TheBobkin · Accepted solution]

Hello Chris,

"I have two all flash vSAN nodes with a separate witness."

If you have only 2 data-nodes then 'Full Data Migration' won't work (unless all your data was FTT=0 and there is neough space on the remaining node)- this is expected behaviour .

Moving the second data-mirror of an Object onto the node that the first data-mirror resides on wouldn't help increase the availability of the data as single node failure would result in loss of both copies and thus the action of moving this data would violate the rules of the applied Storage Policy.

Bob

[ChrisFD2]

Brilliant, thank you. The bit that made it 'click' for me was this:

Moving the second data-mirror of an Object onto the node that the first data-mirror resides on wouldn't help increase the availability of the data

That makes perfect sense - not sure why I didn't realise this sooner!