cluster vsphere HA heartbeats NFS and FC

[speraincorrao]

Hello,

in our infrastructure vSphere 6.7 we have a cluster with 20 nodes. Each node access to the same datastores.

We have 2 kinds of datastore: FC and NFS. (2 Datastore FC and 3 Datastore NFS). All vm's are distribuited in this datastores.

My question (stupid question i know) is when i configure HA heartbeating on the cluster i can choose automatic or Specify preferences for the datastores that vSphere HA uses for datastore heartbeating. Have i to leave automatic or have i to indicate the 5 datastores (2 FC and 3 NFS)?

In this moment we have the HA heartbeating configuration indicating only 2 FC datastore for heartbeating. Have i to add the 3 datastore NFS in our configuration or simple switch to automatic?

In general, If i make a configuration on HA is better to disable and enable HA again or just apply what i've modified?

Sry for my bad english

[bryanvaneeden · Accepted solution]

Hi @speraincorrao,

First of all, there are no stupid questions here! We are all here to help each other. In regards to your questions, please see my points below:

1. It is recommended to leave the Datastore HB setting on "default" which would mean automatic. vSphere HA has a lot of mechanisms in place to choose the best datastore to use for a HB datastore. Think off preferring datastores that are connected on all or the most hosts, datastores that are healthy etc.
2. If you want, you can choose the FC and NFS datastores. It might make the HA HB more "robust", but in my opinion it isn't required.
3. vSphere HA always only chooses 2 datastore for heart beating. So selecting the FC and NFS datastores does not mean it will actually use both).
4. You don't need to enable/reable HA on the cluster if you change the Datastore HB function.

I hope this helps! Please let me know if you have any other questions.

[speraincorrao · Accepted solution]

Perfect! Thanks for the answers, i'll apply what you said.

Just another questions, if you can help.

Consider that we want to dismiss the 2 FC datastore.

I know that:

1. Unregister all objects from the datastore including VMs and Templates
2. Ensure that no 3rd party tools are accessing the datastore
3. Ensure that no vSphere features, such as Storage I/O Control or Storage DRS, are using the device
4. Detach the device from the ESX host; this will also initiate an unmount operation
5. Physically unpresent the LUN from the ESX host using the appropriate array tools
6. Rescan the SAN

So if i put Datastore HB in automatic (default mode) and, after what i mention above, i make a detach of the FC datastore but, for example, in that moment this datastore is used for heartbeats i think that i have no problem (APD/PDL)? Right?

[bryanvaneeden · Accepted solution]

Great. 

Even if detached, the LUN is still present and connected to the host, unless you unpresent it from the Storage Array. I would suggest unmounting it all together -> detaching the LUN and changing the vSphere HA Heartbeat configuration if this is one of the two datastores you are explicitly using for datastore heartbeating. If you have it at automatic, vSphere HA will look for another datastore. This is why the automatic option is preferred. You don't have to worry about it.

[speraincorrao]

Thanks for you time, i'll follow what you said.