This is a straighforward operation. Before removing the shelf and drives from the first system, make sure the drives are no longer part of a volume group or disk pool. Then just follow the documentation for adding a disk shelf to an already existing system.
They already tried it some time ago and it made one of the controllers to panic.
Today, I already scheduled to be on site to check the configuration of disk shelves that will be added. The goal was to verify if any old configuration of DDP or VG was still there. I will have the former E5424 available to reconnect shelves on it and do adequate checks
If so, I will destroy any existing DDP or VG, but the customer says he has already done it. So, at the end of the day, I will be sure that no DDP or VG are present on the shelves they need to add on another array.
My first question: Is there a special command or action that could make me sure shelves are in "factory" defaults and that all configuration (even low-level) has been removed ?
While checking the configuration, I can see that cabling of current shelves does not follow best practices (not a top/down cabling). I suspect this may be the cause of the controller panic when they tried to add shelves.
My second question: Can I change the cabling as a hot operation , or is it mandatory to stop the E5424 array so they can correct the cabling ?
Finally, disk trays that are added contain plenty of drives belonging to the former array. I read that in such case, drives should be inserted one a time. WHat should we do there ?
To make sure the configuration is clear. in the AMW go to Storage Array > Configuration > Clear > Volume or Storage Array. If you are truly trying to set things up fresh, you could just clear the array, otherwise just go ahead and do the volumes to clear the volume and volume group data. Good idea: save the configuration before clearing, just in case.
Remember, E-Series uses what is called a top-down, bottom-up cabling method. You should just need to follow the E-Series cabling guide. Yes, you can do the modifiacation with the array running.
The goal is to correct this configuration that was made years ago.
So if I summarize what should be done:
- On the array which will be stopped, clean configuration (all VG and DDP already deleted) by Storage Array > Configuration > Clear > Volume or Storage Array to be sure Array is clean. (So I guess that it does the same than set storageArray resetConfiguration=true;) then stop array.
- On the production array that stays alive, correct the cabling to implement top-down/bootom-up cabling. This can be done with array live (on FAS systems, a failover/giveback would be recommended, is this the case with E-Series ?)
- Once cabling is correct, add disk trays (which where connected to the stopped array) one by one following usual hot-add procedure. I guess that some disks might be marked as foreign (it was the case when the customer tried it alone). In such case, just initialize the disk. I assume that it is not necessary to add disk drives one by one. Are you OK with this ?
- Expand VG and DDP on production array as needed.
Do you think this should be OK and would not interrupt service ?
Your plan looks good to me except I didn't understand the system with cabling issues is a production system. You can fix it live and you don't need to do anything like a FAS system, but you will affect the production system, and therefore any one using it at the time. I would not recommend fixing it live. If you can, I would suggest taking the system down, fixing the cabling, and then bringing it back up. After it is back up without cabling errors, you can go ahead and hot add the new shelves.