ONTAP Discussions

RDM

alacard052022
391 Views

Hello community. I work in a shop that has Vsphere and Netapp. I am involved in upgrading a to new hardware all the way round. My question current environment uses RDMs... I there a reason why i would want to continue this trend or just convert them to a VMDKs?

 

Thanks!

1 ACCEPTED SOLUTION

Don't recall it ever happening like that in VMware, but i've seen it in other hypervisors.    Usually related to either having way to may paths. (like 1000+ to each host). and incompatibility interoperability across the stack.  i.e. ONTAP was decently currently version, so was the hypervisor, but FC switches were way out of date.  
Have you opened a support case with either NetApp and/or the Hypervisor?  

View solution in original post

4 REPLIES 4

SpindleNinja
374 Views

Few reasons I can think of...

If you're wanting to move away from block,  you need to migrate off RDMs.   

And.. it could also depend on the size of the disk(s).   i.e. how big of a VMDK do you want to manage? 
For example... Back when I was an admin, we had a rule of thumb that anything over 1TB was an RDM and anything less was a VMDK.   

Thanks for the reply. That sounds like a really good plan. I have a few that are at 1tb and a lot that are under by a lot.  Where will still be using block so I think that will work moving forward.  Also in your experience if you have ever been down this road. Had a guest os all of the sudden lose access to RDM. In Disk management  you could see the disk on a offline / unallocated state.  Is there any way to recover it or make the guest os recognize it?

Don't recall it ever happening like that in VMware, but i've seen it in other hypervisors.    Usually related to either having way to may paths. (like 1000+ to each host). and incompatibility interoperability across the stack.  i.e. ONTAP was decently currently version, so was the hypervisor, but FC switches were way out of date.  
Have you opened a support case with either NetApp and/or the Hypervisor?  

TMACMD
367 Views

You could also migrate them to iscsi LUNs connected directly to the vm. 

then if the vm needs to migrate, the iscsi lun will too since it is connected via ip

Public