migrate CIFS+NFS Servers from NSX Celerra to VG8 Celerra - no data migration

If you zone your storage array so that the VG8 can also see your Clariion you should be fine.

NOTE: Your VDM's and shares will need to be re-created on the VG8.

If you have the IP space I would create new vdms, cifs servers and NFS servers then create a migration schedule to unmount the share (file system) fron NSX and re-mount and export from VG8. I would not try and mount the LUN from both at the same time.

As always test your process on a non-production or unimportant share.

No – it is not that simple of just changing the LUNs from the NS-X to the VG8 and mounting the file systems there

If you have the space you could setup the VG8 in parallel and use Replicator to migrate the data to new file systems and LUNs

Please work with your EMC support / professional services to work out a way to do this

Thanks Rainer and Scot for prompt replies.

I understand that it is not as simple as presenting LUNs from NSX to VG8 and mounting there. The configuration of the CIFS and NFS Servers also needs to be present on the VG8. They must share all the file data out with same permissions as previously done on NSX. CIFS must integrate with Active Directory Correctly etc.

However if I look at page 31 of this white paper -> http://www.emc.com/collateral/hardware/white-papers/h8174-migrating-celerra-vnx-replicator-wp.pdf there is an option to migrate VDM between data movers. The context is based around migrating filesystems also however I want to keep the same array and don't have a lot of spare space to migrate filesystems within the same array.

So based on information in white paper is this possible? If not why not? I could for example understand if there were potential blocking issues due to have 2 x CIFS access same LUN but this can be covered by shutting down VDM / CIFS servers on NSX prior to migration and cutover to VG8.

Any more information / discussion much appreciated.

Kevin.

K,

Based on what Rainer has stated and some afterthought this may be more difficult than expected especially when dealing with file systems that span multiple meta's and internal file system structures that would need to be vetted against Dart revisions. There are a lot of variables that would need to be reviewed before even considering this type of move and in the end you still want the result to be an EMC supported migration should you have problems down the road.

At the very least I would engage your CE to discuss. We underwent a similar 23Tb migration but in our case both the Block storage and the Celerra were new.

If I understood you correctly you are not just moving VDMs and file systems between the physical data movers of ONE VNX

You have TWO separate VNX systems then.

It’s even more basic then the fs or meta level – you can’t and shouldn’t have the same LUNs assigned to two completely different VNX

Please work with your local EMC team – what you want to do shouldn’t be done based on comments on a forum