HI, I'm not at the office right now but I had this issue previously.
I think there may be a configuration option you can set on your hosts Configuration Tab | Advances Configuration | (Datastores?)
That can tell a host to ignore that it is a snapshot LUN, not sure about setting it via a script though.
but that advance configuration was for ESX3.5 and not for 4.x
I had this problem with 4.0 servers on a mixed farm, as I say I can't look at it today. just trying to assist.
Has anyone found anything on this?
We appear to be doing something similar to this - but using a SAN snapshot to create a backup of a volume, that contains a VMFS datastore. The backend storage is on a NetApp filer.
What I amn doing to access the data on the snapshot is as follows:-
1. Use NetApp to create a flexclone of the snapshot (this does not take a lot of SAN space)
2. Assign an initiator group the the flexclone so that the VMWare cluster can see the flexcloned datastore
3. Resignature the cloned datastore (this does not affect the original datastore)
4. Import the VMWare Guests
5. Power on the 'duplicate' VMWare Guests, but leave the Guests on the original datastor powered off - VMWare asks if they have been moved or copied; we say moved.
When we no longer need access to the snapshot - we simply remove the Guests on the cloned datastore, remove the datastore, remove the clone (on the NetApp filer) and then we can start up the original ones.
We could have the two sets of guests powerd at the same time, if we move the copies to a different vSwitch in order to prevent IP Conflicts.
This is the only way we have found to do this - but I appreciate we are relying on the NetApp filer not VMWare snapshots so it may not be applicable to your sceanrio.,
we can mount the datastore using force mount command in esx...but we have multiple esx hosts and need power shell script to mount the datastore.
you could do it using invoke-expression and plink.exe
basically it involes using powercli to call the pink command and passing the commands you want to run..
then you could just script for each of the servers your needing to do this on..
to do this though your best using quest rc version of putty as the plink version has some extra parameters to store the key cache on connection.
with the normal version of plink you need to interact with the server to say store the key cache credentials..
Ive attached an example of using this method..
You can use this to send any commands to the host you can do from ssh. and with a little for loop can do it accross all hosts.
Might be helpful.. have a try.
example using plink.ps1 2.0 K
Thanks .......will try this
I have the same problem.
I create with the Data ONTAP PowerShell module a FlexClone of a SnapmirrorVolume, mount the iGroup and switch it online.
After a HBA VMFS rescan the ESXi Host can see the LUN.
In the vCenter Client i can connect this LUN. Now i try to find a way to do this in PowerCLI.
I have the CanonicalName like "naa.60a980005034472f544a6f416d4a487a", VMFSUUID like "4cea85d2-cd86f706-9333-d8d385a3d750" an ConsoleDeviceName like "/vmfs/devices/disks/naa.60a980005034472f544a6f416d4a487a".
Someone an idea to mount the vmfs datastore without format or resignatur?