Does Primary Data to be Snapvaulted to another location have to be in Qtrees?

[davieb1969]

Hi,

I need to SnapVault CIFs and VMware to a secondary location. Can someone provide a definitive statement that source data must be located in Qtrees?

I know for Exchange and SQL when using Protection Manager/SnapManager, data must be in Qtrees, but is this also true for everything else?

The customer already has data located in the root of each volume and I don't want to have to move it.

Thanks,

[sowens]

You have a couple of different options with SnapVault

For the source volume, you can do

/vol/<vol_name>

The above will backup everything in the source volume including source qtrees to the destination qtree. However, since the data is being backed up to a qtree on the snapvault secondary, the source qtrees will not be recognized as a qtree on the destination.

*NOTE the above example is only valid with SnapVault

The second option is to use a source path of

/vol/vol_name/-

The above will backup everything from the source volume that is *NOT* in a qtree.

[adaikkap · Accepted solution]

That's not a requirement for normal dataset in protection Manager.

The specific case of SME and SQL are because non-disruptive lun restore works only for qtrees, and not volumes.

When a entire volume is added to a dataset in PM its replicated entire qtress inside the volume including the data inside the volume and not in any qtree.

So if you have a volume vol1 with 1 qtree qt1, then PM will create 2 qtree snapvault or snapmirror relationship as follows.

/vol/vol1/-

/vol/vol1/qt1

Regards

adai

[davieb1969]

Fantastic Adai, thanks for the helpful information.

dB