Peter Bortas wrote:
>> The way it works is that each dataset (file system)
>> will get a name which is the dataset name with the
>> '/' characters replaced with '_' (underscore)
>> characters.  As Alan explains, you don't get the same
>> directory hierarchy.
> 
> I was afraid of that. I can see why, but it clashes horribly with
 > the "make  lots of filsystems" zfs way, so unfortunately I will have
 > to stick with Samba.

I don't know your situation but with thin provisioning, shared storage
within a pool and the ability to grow zfs pools and datasets, I would
consider encapsulating each project within a file system, and perhaps
use additional file systems to export different views onto each project, 
rather than build projects across multiple file systems.

> Any point in filing an RFE for an extra VFS-layer that works around this?

RFE's are always welcome but it's not a small project due to all the
functional and behavioral consistency issues across mount points.

Alan

>> There is a better description of what happens on my
>> blog at http://blogs.sun.com/dougm
> 
> Thanks.
> 

_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to