vSAN Network Down - HA Action

[virtualizingtheworld]

Hi,

Noticing that when both Host Uplinks dedicated to vSAN traffic go down, HA does not restart the VM's of that Host.

I think this is because the same Host is still sending heartbeats to the connected FC Storage datastore used by HA, is this correct?

Is this the expected and recommended outcome or HA should be configured differently to restart the VMs in such condition since VM's on vSAN have no storage access.

Thanks

[Paudie O'Riordan]

Duncan does a good job explaining things here 

https://www.yellow-bricks.com/2017/11/08/vsphere-ha-heartbeat-datastores-isolation-address-vsan/

 

[virtualizingtheworld]

Ok but not sure how it answers my question, ie. the Host that has lost all vSAN uplinks is still configured to send datastore heartbeats to the FC SAN datastore so HA does not kick in even though the VM's residing on the same Host and on vSAN would go into Read Only mode.

Would enabling HA Host Isolation restart the VM's in such scenario?

[Paudie O'Riordan]

I suspect what's happening is an isolation address needs to be set for the vSAN n/w i.e. you need to use an isolation address that is in the same network as vSAN
I'm paraphrasiong from his blog but..

This way during an isolation event the isolation is validated using the vSAN vmkernel interface, and set the isolation response to power-off. and  if you have traditional storage, then you can enable heartbeat datastores. it will allow the HA hosts to communicate state through the datastore.

here's a table  from his blog post

 

 

 

 

[virtualizingtheworld]

Understood, so setting an Isolation address on the vSAN network such as SVI interface and setting HA Isolation Response to restart VMs (rather than leaving this disabled)

[Paudie O'Riordan]

Yep from your description I think that's what's going on.. 

if you've a lab / pre-prod env, its easily validated

[virtualizingtheworld]

ok thank you