VMware Cloud Community
MichaelLeone
Enthusiast
Enthusiast
Jump to solution

Confused about creating the shared storage for MIcrosoft CLuster in ESXi 5.5

I am looking through the "Setup for Failover CLustering and Microsoft Cluster Service". I am doing cluster across boxes. In the section about adding the shared disk to the first host, it says to add a RDM Disk. Select an unformatted LUN, select a datastore.

How can I have a datastore on an *unformatted* LUN??

Am I supposed to a datastore to my VMware cluster, and then choose this datastore as the location of the RDM disk?

I have created and presented a small 10G LUN from my SAN to my VMware cluster, but I have *not* added it as storage for my ESXi hosts. Am I supposed to?

Very confused. Can anyone shed any light?

1 Solution

Accepted Solutions
MichaelLeone
Enthusiast
Enthusiast
Jump to solution

It's supposed to be a cluster-across-boxes, where each node is on a different ESXi host. Haven't had time to make the anti-affinity rule. Let me move one node to a different host, and try again.

AH HA! This allowed vaildation to work:

"Turn OFF machines, Go to Disk settings then to sharing and change it to Multi Writer on the two VMs and for all disks.

Turn ON VMs, and validate the cluster."

And anti-affinity rules, to keep the nodes apart.

Node:  DCTRTST020.wrk.ads.pha.phila.gov Validated

Node:  DCTRTST021.wrk.ads.pha.phila.gov Validated

Now onto trying to make it into a SQL Cluster. The fun never ends! LOL

Thanks for all the help.

View solution in original post

10 Replies
MichaelLeone
Enthusiast
Enthusiast
Jump to solution

Anyone? Surely it's just something simple I am misunderstanding.

Reply
0 Kudos
MichaelLeone
Enthusiast
Enthusiast
Jump to solution

So I figured out (even though the Guide is seriously vague) that I need to add a 2nd SCSI controller to each node, and set the sharing. I've done that, added my LUN, but am not being prompted to format it in any way. I have added it as an existing HD to the other node, but until I can format the LUN properly, it's next to useless.

Anyone have a better step-by-step than the Guide in the VMware Docs?

Reply
0 Kudos
vijayrana968
Virtuoso
Virtuoso
Jump to solution

When you are adding a RDM, you don't have to format it with with VMFS. The raw LUN is to be formated by guest OS file system of the VM you are adding that LUN (RDM) to. VMFS datastore is used to place RDM mapping file only.

For RDM fundamental, see this VMware vSphere 4 - ESX and vCenter Server

For use case with Windows Server RDM addition step by step, see this detailed post How to create vSphere VMs with RDM disks

Reply
0 Kudos
MichaelLeone
Enthusiast
Enthusiast
Jump to solution

OK! So I've gotten to the point of adding the RDM to the furst node, and added the existing hard disk to the second node.

And I see the (unformatted) disk in DIsk Management in node 1.

SO ... do I initialize and format on node 1? And then add MS Cluster Services? And create a new cluster?

And once I define the cluster, I can import this disk as "Available Storage"?

And once I add the Cluster Services to node 2, and join it to the cluster, it should be able to see the shared storage (presuming that node 2 hosts the cluster role)?

Yes? Have I missed a step?

Reply
0 Kudos
vijayrana968
Virtuoso
Virtuoso
Jump to solution

Correct. Now all steps looks fine. Good luck.

Reply
0 Kudos
MichaelLeone
Enthusiast
Enthusiast
Jump to solution

The cluster won't validate. Disk failures - on the "Validate SCSI-3 Persistent Reservation" test. And I can't install SQL, because the cluster doesn't validate.

Oddly, I was able to move the one disk (which is in "Available Storage", since this will become a SQL cluster, and this disk is meant for it). from one node to the other. That said it moved, and showed no error.

Thoughts? How can I get it to pass validation?

Reply
0 Kudos
vijayrana968
Virtuoso
Virtuoso
Jump to solution

Can you provide exact error or snippet, I guess now that is something need to dig up at OS level.

Reply
0 Kudos
MichaelLeone
Enthusiast
Enthusiast
Jump to solution

Validate SCSI-3 Persistent Reservation

Description: Validate that storage supports the SCSI-3 Persistent Reservation commands.

Start: 11/6/2018 10:55:47 AM.

Validating Test Disk 0 for Persistent Reservation support.

Issuing Persistent Reservation REGISTER AND IGNORE EXISTING for Test Disk 0 from node DCTRTST021.wrk.ads.pha.phila.gov.

Issuing call to Persistent Reservation RESERVE on Test Disk 0 from node DCTRTST021.wrk.ads.pha.phila.gov.

Issuing Persistent Reservation READ RESERVATION on Test Disk 0 from node DCTRTST021.wrk.ads.pha.phila.gov.

Issuing Persistent Reservation REGISTER AND IGNORE EXISTING for Test Disk 0 from node DCTRTST020.wrk.ads.pha.phila.gov.

Issuing call to Persistent Reservation RESERVE on Test Disk 0 from node DCTRTST020.wrk.ads.pha.phila.gov.

Node DCTRTST020.wrk.ads.pha.phila.gov successfully issued call to Persistent Reservation RESERVE for Test Disk 0 which is currently reserved by node DCTRTST021.wrk.ads.pha.phila.gov. This call is expected to fail.

Test Disk 0 does not provide Persistent Reservations support for the mechanisms used by failover clusters. Some storage devices require specific firmware versions or settings to function properly with failover clusters. Please contact your storage administrator or storage vendor to check the configuration of the storage to allow it to function properly with failover clusters.

Stop: 11/6/2018 10:55:48 AM.

Test failed. Please look at the test log for more information.

Nothing in the cluster log itself. The single disk listed as "Available Storage" is an RDM, on a separate SCSI controller (p.17 of the Guide - "NOTE This *MUST* be a *NEW* SCSI controller" - emphasis mine). Disk is set to "Physical compatibility"; SCSI controller is set to share "Physical".

The other node is accessing an existing HD, pointing at this node, also on a separate SCSI controller, also set to share "Physical".

Backend storage is an EMC VNX 5800.

Reply
0 Kudos
vijayrana968
Virtuoso
Virtuoso
Jump to solution

I see, are both cluster VMs on same ESXi host? if that is the case please change compatibility mode to 'virtual' on SCSI controller of both VMs.

You need to set mode to Physical only when both cluster VMs are on different ESXI hosts which is call (CAB) cluster across box.

Reply
0 Kudos
MichaelLeone
Enthusiast
Enthusiast
Jump to solution

It's supposed to be a cluster-across-boxes, where each node is on a different ESXi host. Haven't had time to make the anti-affinity rule. Let me move one node to a different host, and try again.

AH HA! This allowed vaildation to work:

"Turn OFF machines, Go to Disk settings then to sharing and change it to Multi Writer on the two VMs and for all disks.

Turn ON VMs, and validate the cluster."

And anti-affinity rules, to keep the nodes apart.

Node:  DCTRTST020.wrk.ads.pha.phila.gov Validated

Node:  DCTRTST021.wrk.ads.pha.phila.gov Validated

Now onto trying to make it into a SQL Cluster. The fun never ends! LOL

Thanks for all the help.