volume shadow copy with recursive snapshot and a single share possible?

[CrazyKay]

Hello everybody,

voloume shadow copy is working fine at my freenas installation. Now I'm trying to optimize the settings and shares. My goal is something like this:

volume1/projects/
volume1/projects/ProjA
volume1/projects/ProjB
volume1/projects/ProjC
...
and so on where volume1 is the name of the volume and projects is the first dataset. ProjA and so on are further datasets under projects. Then I want to setup periodic snapshot task at volume1/projects/with recursive snapshots and a CIFS share with volume1/projects as path.

Volume shadow copies are working fine if I create a share for every dataset ProjA, ProjB and so on but not if I create a share with the path volume1/projects/. It would be an enormous effort if I have to create a share for every dataset and then assign them to the user.

Is it even possible what I'm trying to implemt?

 

[anodos]

Hello everybody,

voloume shadow copy is working fine at my freenas installation. Now I'm trying to optimize the settings and shares. My goal is something like this:

volume1/projects/
volume1/projects/ProjA
volume1/projects/ProjB
volume1/projects/ProjC
...
and so on where volume1 is the name of the volume and projects is the first dataset. ProjA and so on are further datasets under projects. Then I want to setup periodic snapshot task at volume1/projects/with recursive snapshots and a CIFS share with volume1/projects as path.

Volume shadow copies are working fine if I create a share for every dataset ProjA, ProjB and so on but not if I create a share with the path volume1/projects/. It would be an enormous effort if I have to create a share for every dataset and then assign them to the user.

Is it even possible what I'm trying to implemt?

You might try the share-level auxiliary parameter "shadow:snapdirseverywhere = yes"

edit: which I just realized is a default parameter. So sorry there.

 

[Sakuru]

I'm curious, why are ProjA etc. datasets instead of folders?

 

[CrazyKay]

Because of the capabilities of datasets (quota, permissions, etc)

 

[anodos]

Because of the capabilities of datasets (quota, permissions, etc)

Permissions can be set at a file / folder level, and if this is a business environment then you would be well advised to study all materials on samba permissions carefully.

 

[CrazyKay]

Permissions can be set at a file / folder level, and if this is a business environment then you would be well advised to study all materials on samba permissions carefully.

As mentioned it's not just because of the permissions. It's also because of feautures like quota, snapshot, etc.