DS4700 and adding a new exp810


ADSM.ORG Senior Member
Jun 7, 2004
Reaction score
Great White North

We have an IBM DS4700 with 2 EXP810 already configured and up and running. We need to install a new EXP810 enclosure to that setup.

The original config was done as per the IBM Installation Book (see pic 1) for one ds4700 with 2 exp810. If i am to add a new exp810, the cabling has to change a bit (see pic 2) and i was wondering if i could do this safely, with no impacts? TECHNICALLY, the loops are redundant...but about past experience from some of you guys?

You should be alright. If you wanted to lean towards the extra-safe side, pick a time to do it during downtime, but I would feel safe doing it mid-operation.

I maintain their DS4800 equipment (an Engenio 6998 whereas the DS4700 is a 3994, I think) and have come to feel very confident in the platform.

Since you have dual-paths between your controllers, though, you should be able to do this.
Well, recabling was a breeze. I wish i had a test environment but hey, i did it on our production environment...As long as you do one loop at a time, there are no impact.

But, one more question then, now that my exp810 is attached correctly, i need to insert the drives. By the installation PDF, i need to run the script to Disable Migration, insert the drives then run the script Enable Migration...
Do i NEED to run the scripts? They are brand new drives and from my understanding, these scripts handle the fact that if data is found on the drives, it won't be migrated...But if they are new, is that a step we can skip? Any tips?

DISABLEDRIVEMIGRATION.scr is a best practice approach to and is highly recommended.

The DS4000's save array information on the individual drives allowing for ease of migration between systems (ds4300 to ds4800 upgrades, etc). Because of this feature, you now have the risk of inserting a "new" drive that could have been apart of previous array (ie array1). If you do not run that command, and insert that drive... the drive will automatically assign itself into "array1" causing you array to potentially malfunction or become corruption.

This has happened in the past, however unlikely, I always use the script to avoid being the lastest example of why to run that script.