2 Node Hyperconverged Server 2019 Heartbeat Failover

[suntech.tt]

I want to use the heartbeat failover strategy, however, I'm not seeing much information on some of the questions below
1. Do I still need to set up a witness disk?
I see in some older documentation and manuals that the witness disk is setup on node 1 and replicated to node 2 so I've done this.

2. If Q1 is yes, how do I configure the iSCSI target. This is the confusing part for me because, in some documents, the witness disk has the multi-path check box checked and the failover option selected for MPIO devices and other documents say, do not check the multi-path checkbox and do not connect the partner node to the witness target...

Setup details
Node 1+2 {
OS: Server 2019,
Starwind Feature Installed: Service,
Storage: All Flash,
RAID: Hardware RAID
Windows Features and Roles: Hyper-V, Failover Cluster, File Manager, MPIO
}

Thanks

[yaroslav (staff)]

1. Do I still need to set up a witness disk?

Yes. Witness is needed for a failover cluster. See more at https://docs.microsoft.com/en-us/window ... ter-quorum.

2. If Q1 is yes, how do I configure the iSCSI target. This is the confusing part for me because, in some documents, the witness disk has the multi-path check box checked and the failover option selected for MPIO devices and other documents say, do not check the multi-path checkbox and do not connect the partner node to the witness target...

If you are running a compute-and-storage-separated setup (i.e., you have VMs and storage on different servers; one server for compute and 2 for storage), you need to connect partner's witness over iSCSI. See more at https://www.starwindsoftware.com/resour ... rver-2016/
If you are running a hyperconverged setup (i.e., VMs are running on the same box as storage), you should not connect partners' witness disk and Failover Only MPIO policy should be used. See more at https://www.starwindsoftware.com/resour ... rver-2016/.

[suntech.tt]

yaroslav, I'm running a hyperconverged setup. So these are the steps I've taken. Please correct me if I'm wrong.
1. add device and witness virtual disk to node 1
2. replicate to node 2 using heartbeat failover strategy
3. open iscsi initiator properties on node 1
4. connect the local witness disk as per documentation (enable multi-path ???, select local adapter and target portal [127.0.0.1])
5. leave the partner witness disk "Inactive"
6. open Devices, open MPIO..., select Fail Over Only
7. bring witness disk online and initialize in disk manager. Do not give a drive letter.???
8. open iscsi initiator properties on node 2
9. repeat steps 4 to 7 above ????

[yaroslav (staff)]

Yup, your procedure is right.