Great!!! I would use larger file-systems but I am concerned about file-system corruption and the amount of time required to recover.... and thoughts on that?

you got it.

so for example if you need one mount point you could:

/departments
   /HR
   /Marketing
   /Management

here HR, Marketing and Management are unique file systems that are mounted as component file systems underneath mount point departments.

yes, a full file system check is pretty rare these days and its a lot faster than it used to be

I got the impression that for the time required the number of files and ACLs make much more of a difference than the file system size itself

yes - thats most usefull for CIFS

for NFS its even easier to just mount them like this on the client using automount

Just remember that each component file system will look like a subdirectory and still be limited to its size.
So if one of your 500GB fs is full you cant write into that dir any longer even if the others have space available

what application are you using for backups? There are some applications that will allow you to create "virtual" file systems where you can backup one nested file system using multiple backup streams. For example you could backup /HR, /Marketing and /Management as if there were separate file system ..so you could utilize multiple tape drives/disk pool.

datamovers use journal file system so typically if datamover panics ..it's backup in a few minutes, it depends how many file systems had to be fsck . Typically if i manually run fsck on a 2 terabyte file system that's 80% full ..it takes about 6-7 minutes, but i have never seen datamovers take that long even after they panicked.

document named "Managing Celerra Volumes and File systems Manually" should have what you need. So far i have see that in order to apply tree quotas to nested file systems, you have to umount component file systems, apply tree quotas settings and then remount them again under the root file system.

you can do that ..you are going to be limited only by the limit of how many file systems allowed on Data Mover.

You guys rock.... I appreciate the help.. I'm going to take a look at this in more detail. While I got you... do you have any experience with Rainfinity GNA and/or GFV?

yes, its explained in detail in the "Managing Celerra Volumes and File Systems Manually"

Its on the doc CD and Powerlink in the Support->Documentation section

drop us a note if you dont find it

what are you trying to achieve ? file systems larger than 16TB ?

so.... /departments
/HR
/Marketing
/Management
all could be individual 500gb File systems and the apps could just mount /departments and have access to all?

Yes, I am currently using F5/Acopia to virtualize 120x 500gb file systems into a single mount point for some app. servers. Unfortunatley Acopia doesn't scale very easily , very expensive. so I want to create many file systems and with a single mount on the Celerra.... is this possible?

yes, understood.... that's where Acopia is nice.. it automigrates data to available file systems....

I have 3 active data movers..... do I then need 3 nested file systems?