I am going to build a Windows Server 2012 R2 Failover Cluster test environment on Vsphere 5.5 U1 as a single ESX box with two VM's clustered, I am using SAN storage and RDM disks mapping to VM's directly.
Can anyone share your experience does it supported and will work fine, or there is any special configuration required for 2012 R2, I was trying to find a technical article on VMware but no vain on 2012 R2 support.
Thanks and appreciate all your suggestions..
Per KB article, Windows 2012 R2 Failover Cluster is not officially supported VMware KB: MSCS support enhancements in vSphere 5.5
I does not mean it won't work though Server 2012 R2 Guest Cluster support
As per ESXi 5.5 U1 release notes, this is not supported officially. But it may work.
Release notes:VMware ESXi 5.5 Update 1 Release Notes
if found useful, plz mark as correct or useful
Scroll down to
and it shows
|Supported||Windows Server 20031|
Windows Server 2008
Windows Server 2008 R2
Windows Server 20122
Windows Server 2012 R25
|4.x/5.x||See additional considerations|
Then if you read the notes they show:
C:drive) virtual disks can be on local VMFS or SAN-based VMFS datastores only, regardless of the underlying protocol. System disk virtual disks must be created with the EagerZeroedThick format. For more information, see the Setup for Failover Clustering and Microsoft Cluster Service guide for ESXi 5.0 or ESXi/ESX 4.x.
For support information on Microsoft clustering for MSCS, SQL, and Exchange, go to the Windows Server Catalog and select the appropriate dropdown.
Windows Server 2012 failover clustering is not supported with ESXi-provided shared storage (such as RDMs or virtual disks) in vSphere 5.1 and earlier. For more information, see the Miscellaneous Issues section of the vSphere 5.1 Release Notes. VMware vSphere 5.5 provides complete support for Windows 2012 failover clustering.
VMware vSphere 5.5 Update 1 provides complete support for Windows 2012 R2 failover clustering.
So, sounds like we're getting mixed signals here but according to this article it says it is supported !?!
I believe KB 1037959 is the latest update and from the looks of it, support has been added in 5.5 U1
Looking at the KB's update history: 04/30/2014 - mk - Added support for Windows 2012-R2 clustering with shared storage on 5.5_U1
support was added on 4/30 and we replied to your posts 3 days earlier (when it is still unsupported )
Nowadays I have been struggling with Windows Clustering on ESXi 5.5. Here is my environment;
2 ESXi 5.5 Update 1 hosts.
2 Windows 2012R2 Servers; each one runs on a different host (I set seperate them by using DRS affinity rules).
All system disks are ZeroLazyEager.
I have 5 shared disks (VMDKs) and all of them are set as ZeroLazyEager.
All virtual SCSI controllers are LSI Logic SAS.
Windows Cluster looks like running properly but it doesn't let me install clustered MS SQL 2012. MS SQL 2012 installation complains that my disks are not configured properly.
Windows cluster can not validate "Validate SCSI-3 Persistent Reservation".
In spite of documentation, ESXi 5.5 U1 doesn't support MCS yet or I miss something.
Need to your advises...
Like Rommel Posted it looks like they updated the KB article to include it, however the arguement for MSCS essentially always burrows down to "Patching" now since the Vmware Cluster has all the fault tolerance you could want. In reality MCSC doesn't failover seamlessly either. In most cases it can take anywhere from 5s-1+min, depending on the cluster and how much stuff it has to failover. I have seen large file servers / SQL clusters take upwards of 2-3 min to failover. Now in most cases the users are unawear of this or if they get an error they try again in a min or two and everything is fine, however the rest of a VM is 1-3min tops now-a-days, even with SQL. With this and snapshoting I'm seeing many people just remove there MSCS completely unless they stay on physical servers.
The administrative overhead of MSCS inside VMware is just not worth it IMO and I fight it to the bitter end, needless to say I have lost that battle a few times, but you win some you loose some.
Currently in most cases right now when we setup patch cycles I have a powerCLI script that runs and snapshots everything that is being patched by WSUS prior to the patch cycle, then if everything is good after the patch cycle, I run a script to remove all the snapshots.
If you have vSphere Enterprise licensing with VCOPs with Config Manager this is done for you automatically.
Also FT with more then 1vCPU is coming down the pipe, I have hopes for version 6 so lets keep our figured crossed as this could eliminated MSCS inside VMware completely
Hope this has helped
We tried the same:
2 hosts with ESXi 5.5 Patch 4 (2403361) using HA and DRS (DRS rules to ensure MSCS cluster node separation)
FC-attached Storage (Fujitsu Eternus DX200)
Two identical MSCS cluster nodes:
We also got validation errors regarding SCSI-3 Persistent Reservations
I resolved the issue today by deleting and re-creating the RDM files and setting the RDM compatibility mode to physical (we had virtual mode before)
If you think about installing windows server updates on the sql nodes you sometimes need some more time than just a few minutes.
Especially if you run some applications that freak out completely when a sql server is going offline a MSCS inside VMware is definitely useful.
In everyday operation a MSCS is not much administrative overhead imho but when you need to take a VM offline it's nice when you can just do a failover.
Just my thoughts