I have confiured SRM for other VM servers but I am having trouble with one that has RDMs connected.
I originally set up the VM to use SRM and detached the RDM. I know want to have the RDM connect automatically in a DR situation.
What settings and configuration do I need to for:
a. The lun to be seen at the disaster recovery site?
b. The lun to be seen by the SRM configuration?
I cannot locate any documentation on this....
FYI, my setup is as follows:
ESXi 4.0 hosts
NetApp Data Filers
I can't get u clearly... may I know what you mentioned as other VM servers? I understood like the VM servers which are not have RDM disks. So are you succeed with the configuration the VM servers without RDM? meaning like r u created recovery plan for the VMs which dont have RDM and executed those recovery plans?
So long as your VM, and the RDM LUN are replicated, you should not have to do anything special. Since almost 1.0 RDM have been fully supported.
If you do have issues, sometimes it is due to the RDM LUN not being replicated.
But if you are replicating both the VM, and the RDM, and they are replicated by the same NetApp you should be fine. If not, you will need to provide error messages and logs and you should always think about using our support.
I have previously successfully set up other VMs to work with RDM. They have been tested and failover to the recovery site properly. These VMs do NOT have RDM luns attached.
The VM that I have which has an RDM attached I cannot setup the SRM for it properly.
when you mention cannot, can you please tell us the error message?
There is no error message. I am unable to set it up because I don’t know how to do it. There is no process or documentation advising how to do this.
You can just follow the step what are all for the VMs which dont have RDM disks. There is no difference to follow the steps for RDM disks. Just try the steps and let us know if u face any issue in that..
But it does not work in the same way as the non-RDM servers as there is not a datastore at the recovery site to connect to…
So i think your VM's guest OS in RDM disk. While creating VM the configurations will be saved under one Datastore. You can use that Datastore.
As mentioned before, SRM fully supports protection of VMs with RDMs. So, you handle it as regular VM.
What you need to do is:
1) You should establish replication of RDM LUN, exactly as you established it for LUNs backing your VMFS datastores. No any special configurations needed.
2) I understand from your post that currently your RDM is not attached to VM. Attach it (RDM mapping file must reside on replicated datastore).
3) If replication is configured correctly, you will see LUN backing RDM in array managers (in the same datastore group that VM's VMFS datastore LUN resides in).
4) VM's ststus in Protection Group should be "Not Configured". Right click on VM and choose Configure Protection. That's all.
SRM will automatically attach the RDM to recovered VM during test or failover.
Is your NetApp Filer an active/active cluster, or do you only have a single controller? I am running into the same issue, but at first, my RDMs were located on a different controller than the location of the RDM mapping files. The mapping files for the RDMs reside with the VM which was on controller A, while the actual RDM LUNs resided on controller B. I created a new LUN on controller B in the same volume as the RDM LUNs and attached it to ESX as a new VMFS datastore. I then migrated the VM to that store so that the VM and the RDMs were being hosted by the same controller. Waiting for a new snapmirror to be created for that volume, then will attempt to do this again. From my research, if you have RDMs on different controllers than the actual mapping files, the array manager does not understand how to map the files to the actual RDMs. When the snap mirror is finished and I attempt to configure the array manager again I will post my results.
production: NetApp 3140 cluster, active/active
recovery site: NetApp 3140 cluster, active/active
ESX 4.1 U1
Exchange 2003 VM, OS installed on VMFS partition, Database and Logs installed in physical mode RDMs.
All LUNs for this VM (RDM and VMFS) are located on same Volume controlled by controller A, volume size 2TB, connected via iSCSI
Snapmirror relationship for volume between production and D/R site. Snapmirror currently initializing
Ok, so my theory was proved correct. I created a new LUN on the same volume as the rest of the RDMs, created a VMFS partition on that LUN and moved the Exchange server VMDK to that datastore. Snapmirrored the volume, then ran the array configuration in SRM from the source side. The new datastore appeared, along with all the RDMs associated on that volume. I was then able to enable protection on the Exchange server. BTW, these are physical mode RDMs, not virtual mode. See the pic attached
Is there a KB that details the steps to SRM a VM that has RDMs?
There isn't. The steps are the same as for a VM that doesn't have an RDM (see above)