VMware vMotion clustered RDM?

EE, thanks in advance for reading

We have a Microsoft cluster VM with RDMs in Physical mode for the controller type. We get the error below when trying to vMotion to another host. (Storage located in a DRS cluster)

"Device 'SCSI controller 1' (&2) is a SCSI controller engaged in bus-sharing"

The VMware article states the workaround is to turn off the server and migrate or to switch the mode from Physical to Virtual but that also requires a shutdown. The latter method actually isn't really cluster supported since shared machines must be pinned to the same host thus defeating the purpose of clustering and highly available, correct me if I'm wrong.

Two questions,

1) Is there another workaround to migrate the VM live maybe via command line i.e. convert the controller to virtual or none then switch back as soon as the machine is migrated?

2) Why use M$ cluster in VMware environments with Fault Tolerance in 6.0? Additionally, you have shadow copies, DFS and SAN level snapshots in addition to your backups and keep it simple.

Who is Participating?

[Product update] Infrastructure Analysis Tool is now available with Business Accounts.Learn More

I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Is this Cluster in a Box, e.g. two virtual machines using a VMDK for storage ?

are both VMs on the same host at present ?
Mr TorturSystem EngineerCommented:

1) no, as far as I know
Maybe - a shot in the dark - you could shutdown one of the two VM, detach the RDM, VMotion the VM and reattach the RDM. And then the same if you want to VMotion the other VM. That's what I would try, but don't do it without verifying there would be no problem detaching/reattaching RDM.
Also I don't know about M$ cluster specifically so you have to manage the operation at the M$ cluster level too.

2) in fact none, unless you really need a M$ cluster. In fact in the vsphere 6 version they really have enhanced (is it correct in english? ;-) well sorry if not) the Faul Tolerance feature, going from 1 vcpu max in 5.5 to 4 vcpu max now. Also now 64 GB RAM max, FT support thin disk now too. And it is said to be womehow "vsan supported", that's what a teacher told me in a vsphere 6 course but don't remember what that was about.

You should answer Andrew, generally he's not the last to have an idea  ;-)
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
a bit OT, but vSAN 6.1 (note the release does now support FT).
IT Pros Agree: AI and Machine Learning Key

We’d all like to think our company’s data is well protected, but when you ask IT professionals they admit the data probably is not as safe as it could be.

Mr TorturSystem EngineerCommented:
Hi Andrew,
OT? Just to understand, what do you mean?
snyderkvAuthor Commented:
We aren't using vSAN as I thought that was for environments without SAN storage.

Andrew, we are using clustering via pRDM so that the each shared VM can be on a different host. I want to present a change to use CIB (cluster in a box) so we can convert to VMDK and migrate them off hosts. Well use a DRS rule to pin both to a host. There is no SLA agreement to have redundant redundancy as a Microsoft Cluster should be enough.
snyderkvAuthor Commented:
Mr Torture, he means off topic most likely.

By the way Andrew, nobody mentioned but I think 6.0 now supports vMotion of pRDMs within a Microsoft cluster unless I'm mistaken.
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Okay back yo question, so you have physical RDMs.

You want to use Cluster in a Box, wow - this is only really designed for Labs, and Demos, not for serious production! Because it cannot be vMotioned, and you will need to shut it down, to move off a host, when you need to do maintenance, and it becomes a problem to migrate! (even off!) - and you'll need to do manually.

CIB is not a good idea.

That's correct OT - Off Topic. vSAN is for anything you want to use requires extremely fast storage, which is included and designed into vSphere.
snyderkvAuthor Commented:
Andrew, vSAN requires a spindle disk and 3 drives locally installed plus three hosts?. Those are odd requirements. We use blades with only two drives so no can do I believe.

With CIB you can vMotion Microsoft Cluster VMDKs "live". I'm not sure what you're talking about in your last statement.

And not sure why it's only for lab environments. How do you think people clustered with physical hardware for the last couple decades? They had no host at all, just two servers, clustered. CIB is simply that setup but attached to a host.
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Four hosts recommended for vSAN (although 3 will do, and now there is a ROBO 2 host version), and now all can be flash, or hybrid flash. (HDD & SSD, or PCI-E Flash)

Blades would not be suitable, purchase vSAN Ready Nodes, or new servers, with disks losts, of EVO:RAIL 4 ESXi nodes in a box.

I was referring to storage vMotion.

We are talking about Cluster in a Box, the sharing of virtual machine disks (VMDKS) between virtual machines on the same host, using virtual scsi bus sharing?

Most use Cluster in a Box for Test and Dev not production, or demos, experimentation, proof of concept not production.

Physical machines, use shared storage, e.g. LUNs. (either RDM) or iSCSI LUNs.

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
Mr TorturSystem EngineerCommented:
Ok for OT, thx.
vSAN requires at least 3 hosts with at least 1 SSD and 1 HDD each.
As it is built on internal drives, you can do it without a San which could be a big benefit.
And as the design is to have disk redundancy bewteen separate hosts, a vSAN on blades within the same chassis would be a bad idea IMHO.
snyderkvAuthor Commented:
Tortur, gotcha on the blade, I was never planning on it but good to know

pRDMs with SCSI bus sharing set to Physical. So the Microsoft cluster behaves like a physical cluster. The server operating system is still HA only it's attached to one host instead of two, but at least you can vMotion it once you convert the disks to VMDK. I'm not sure why the cluster has to be split among separate hosts when the operating system is still clustered.
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today

From novice to tech pro — start learning today.