How to exclude VM from datastore snap?

[#1]

I have 2 NFS Datastores and 2 snap jobs in my SMVI 2.0, each one for each datastore.

One of the datastore has 3 VMs that are not changing (archive versions), so how can i exclude them from backup, and still back uo every thing else on that datastore?

[#2]

Hi,

I'm not a smvi specialist,but because SMVI at the end just make snapshots of the datastore, I can't see how you can exclude some VM's of that snapshot. It's not possible to exclude files (or VMDK's) from a snapshot.

But why do you want to remove it from the snapshot? When it not change, it doesn't use snapshot space.

Reinoud

[#3]

I want to exclude it because I'm backing up these snaps to tape every night, for long term archive and don't want these static machines to take up space on tape every nighjt.

I heard somewhere that if I mark vm drive as independent/persistent it will be skipped during smvi snap, isn't it true?

[#4]

On the SMVI level (but I don't use the product) it's probably possible that you can exclude those VM's (and disable the pre and/or post processing on the VMWare level for those VM"s). But because they on the same volume as other VM's, they will always be on the volume and include in the snapshots of those volumes.

If you every night take a full dump to tape from those snapshots, the static vmdk's will always in those snapshots. The same if you put some files directly in that volume.

Unless I miss some magic from SMVI, I think that the only solution is to put those static VM's in an other volume.

Maybe a SMVI expert/user can confirm this?

Reinoud

[#5]

But I don't back up entire volume to tape, just the most recent smvi snap directory, that's why I would like to exclude these vm's.

Anyway, I will try to set the hdd as independent/persistent and we will see if smvi skips it.

I will let you know.

[#6]

Reinoud is correct.

After all of the orchestration of complete, SMVI ultimately takes a NetApp Snapshot of the Volume (aka datastore).  The only way to remove certain VMs from the tape dump of that volume is to remove those VMs from the datastore.

Many of my customers create datastores (volumes) based on the data protection (or SLA) required for those VMs.

Datastore_Gold = (hourly snapshots + snapmirror and backup to tape)

Datastore_Bronze = (daily snapshots)

Cheers, Tony

[#7]

That VM will exist in every snapshot of that volume. WAFL will determine the space usage based on the delta between snapshots. If you are doing file based backups (like NDMP) and trying to be space conscious there, you will need to rely on that mechanism to exclude that object.

The independent, persistent disk setting only prevents chaos by keeping VMware from forcing other disparate datastores to take snapshots when a VM spans more than one datastore. Just imagine if these other datastores had VM's that spanned multiple disparate datastores.

Please pardon my BlackBerry thumbmanship ...

Stetson M. Webster

Professional Services Consultant

NCIE-SAN, NCIE-B&R, SCSN-E, VCP

NetApp Professional Services - East

919.250.0052 Mobile

Stetson.Webster@netapp.com

Learn how: netapp.com/guarantee

[#8]

It is  recommended to  create a datastore dedicated to transient and temporary data for all virtual machines with no other data types or VMDKs residing upon it. This is done in order to avoid a Snapshot copy being performed on the underlying volume as part of the backup of another virtual machine. It is recommended to exclude datastores that contain transient and temporary data from the backup, this ensures that snapshot space is not wasted on transient data with a high rate of change. In SMVI 2.0, when selected entities in the backup span multiple datastores, one or more of the spanning datastores may be excluded from the backup. Once configured, the transient and temporary data VMDKs will be excluded from both the VMware vCenter snapshot and the NetApp Snapshot copy initiated by SnapManager for Virtual Infrastructure.

Regards

Amrita