I posted something like this last year but didnt get a lot of feedback, so I'll do it again.

I'd really like to get any opinions on this.  Feel free to use this in your own environment.

Every year our company phyically breaks our network to our DR site and performs a actual DR test.  We have a AD Domain controller at that site.

During that time we bring the up the CIFS servers on the target array, leaving the file systems as R/O.  Last year was our first year in doing so and after a few glitches we were up and running.

The glitches were a corrupt vdm.cfg file (but thats another story)

Here you go, please respond.

STEPS TO DISCONNECT NETWORKS

On the TARGET

Open Putty Session

server_ifconfig server_2 replication down

On SOURCE: NS40

Open Putty Session

server_ifconfig server_2 replication down

Open Celerra Manager, the Replication sessions should show an error state indicating network communication lost. 

On the SOURCE: NS40

Stop all replications going to DR site.

Open Celerra Manager GUI

Select Replications

Sort on Data Mover Interconnect or Celerra Network Server

The replications that are in the column of Celerra Network Server that are the TARGET are the ones that need to be stopped.

Select STOP, and then select on Source ONLY

Select OK

Replication sessions should now be in STOPPED state

Once communications have been lost and all replications stopped, it will be safe to “break” the connections between Source and Target Site.

Contact Data Network Team and inform them it is now safe to break the connection between sites.

Once physical connectivity between sites has been completed and verified, continue to set the correct states of the VDM  resources on TARGET

On the TARGET:

Change the VDM(s) state from “Mounted to “Loaded”

Open Putty Session

Be sure to get a new list of VDM’s in the event this documentation has not been updated.  To do so, run this command.

nas_server –vdm –list

Change the VDM(s) state from “MOUNTED” to “LOADED”

nas_server  -vdm prod –setstate loaded

Ensure all File Systems are R/O at the TARGET to ensure no files are edited on the TARGET file system during the DR tests.

On TARGET:

Open Putty Session

Run command to get a full list of the File Systems as this list may not be updated.

Nas_fs –list         

Changes to DNS will be required at DR Site in order to reflect the IP addresses in that location.

Contact Data Network Team

Change DNS entry of CIFS_00 to – 10.0.0.1

Change DNS entry of CIFS_01 to – 10.0.0.2

etc

Once the IP addresses have changed in the TEST environment AND the Domain is up and running along with the DFS roots being available, all DFS links should be available.

\\dfs\share1

\\dfs\share2

\\dfs\share3

DR TESTING BEGINS

*********************************************

DR TESTING CONCLUDES AND RESTORATION IS READY TO BEGIN

Restore back to Production / Target Use

ENSURE THE NETWORK BETWEEN SITE IS STILL DOWN

Remove the CNAMES in DNS

Contact Data Networking Team at DR site to ensure the IP addresses for the CIFS servers are set back to original IP addresses for the DR site.

Set TARGET VDM state to mounted

On the Destination:

Open Putty Session

nas_server  -vdm prod –setstate mounted server_2

On the Destination:

Open Putty Session

server_ifconfig server_2 replication up

Bring up the interface on the source

On Source:

Open Putty Session

server_ifconfig server_2 replication up

CONTACT DATA NETWORKING TEAM TO BRING NETWORK BACK ONLINE

Startup Replications

On Source:

Open Putty Session

VDM

nas_replicate -start prod_vdm_rep -overwrite_destination

FILE SYSTEMS

NOTE:  Do not start all of these replications at once.  They must be done one at a time and will consume most of the next several days in order to get up to date.

nas_replicate -start prod1_fs_rep -overwrite_destination

nas_replicate -start prod2_fs_rep -overwrite_destination

etc

File systems on SOURCE should be RW and the VDM state as LOADED and ACTIVE

nas_server –info –vdm vdmname

File systems on TARGET should now be RO and VDM state as MOUNTED

nas_server –info –vdm vdmname

Monitor the replications resumed on the NS40 to NS20 replications.

Nas_replicate –info replication_session_name

nas_replicate –info prod1_fs_rep