RDM unable to add to second VM if once added to a VM

[Wajeeh]

Dear Experts,

In our environment of Esxi 5.0 U2, on one physical Esxi host, I created two virtual machines of Win 2008R2, I'm not looking for any MSCS clustering. I created a LUN and presented it to host, the size of Lun is 10GB. I added it to VM1 as RDM (virtual) and it is perfectly ok, I am able to see it in windows. When I attempted to add the same as RDM to VM2, I found the RDM choice is grayed out, why is that, is this not possible to allow 2 virtual machines to see same RDM ?

I've attached screenshots for understanding how it is looking like. SCSI bus sharing on VM_1, I selected virtual to share a 'vdisk' between VM's in same host, but that didn't helped either, later I realized it is for sharing a 'vdisk' and I am looking for using same RDM in two VM's

Please see what could be the reason and if this is possible or not, if not possible what are technical reasons ?

waiting for response.

Thanks,

[admin]

Hi,

yes you can add the same RDM as shared between two VM's. But on the second VM you need to add the RDM pointer file and not the RDM directly again.

Check the below link

Adept Techie: Shared RDM LUN between non-clustered vmware virtual machines

Thanks,

Avinash

[ScreamingSilenc]

Change the SCSI Controller type to LSI Logic SAS instead of Parallel and instead of adding the RDM on second VM try adding the Existing HardDisk (_1.vmdk) which is the pointer disk to the RDM.

[Wajeeh]

If I add on second VM as existing hardDisk, are there chances of corruption of data ? is this the right way to share RDM between two virtual machines ?

[ScreamingSilenc]

My bad since your not using any clustering solution this would corrupt the disk as there would be no reservation on the disk by any VM.

[Wajeeh]

Hello Mustafa, In your earlier reply to me you advised me to 'use existing disk' and attach to second VM, now with your reply you are mentioning it will corrupt the disk.

@ avinash, this is what I want to be sure before performing this to environment.

Any other thoughts

[continuum]

You should be happy that ESXi prevents this kind of stunts.

The referenced blog post does not apply to your case - those Oracle VMs sure did not use NTFS.
I would expect that your original plan - if you implemented it with work arounds would corrupt the NTFS in the first one hour of usage.

Use regular Windows-fileshares instead - or create a Windows Cluster.
Dont try to read and write to a vmdk from two VMs when the FS is NTFS.

The referenced blog should have a fat warning that it only applies to the specified special case

[ScreamingSilenc · Accepted solution]

Hello Mustafa, In your earlier reply to me you advised me to 'use existing disk' and attach to second VM, now with your reply you are mentioning it will corrupt the disk.

Sorry Initially I though your using some clustering software that's why I have advised you to use existing disk. Since your not using any clustering software this will corrupt the disk .