IBM Serveraid 4LX, trying to defunc drive from mirror pair
Hello
I have rather a sticky problem. A new customer has an IBM server with two mirror pairs under a serveraid controller 4Lx. The primary OS drives as well as the data are mirrored. The OS drive array has a drive that is constantly rebuilding, it has been like this for nearly two weeks, so it is pretty certain that the drive is actually faulty. With this in mind new drives were purchased.
Within the serveriad manager, it seems not to want to allow me to defunct the faulty drive, and it warns that the array will be taken offline (no good) As I do not know IBM servers or Serveraid very well you can understand my concern. I had assume I could just add a drive as a hotswop, defunct the faulty one and away it would all go?
Backing up this fulty arry is a none even as well, so far it reckons it will take 25 hours! for 130GB, so you can see that the alleged rebuilding is killing the performance of the server.
Anyone know this controller? and can explain how to swop out the drive without taking the other mirror offline?
Regards
I have rather a sticky problem. A new customer has an IBM server with two mirror pairs under a serveraid controller 4Lx. The primary OS drives as well as the data are mirrored. The OS drive array has a drive that is constantly rebuilding, it has been like this for nearly two weeks, so it is pretty certain that the drive is actually faulty. With this in mind new drives were purchased.
Within the serveriad manager, it seems not to want to allow me to defunct the faulty drive, and it warns that the array will be taken offline (no good) As I do not know IBM servers or Serveraid very well you can understand my concern. I had assume I could just add a drive as a hotswop, defunct the faulty one and away it would all go?
Backing up this fulty arry is a none even as well, so far it reckons it will take 25 hours! for 130GB, so you can see that the alleged rebuilding is killing the performance of the server.
Anyone know this controller? and can explain how to swop out the drive without taking the other mirror offline?
Regards
ASKER
Well yes, that was my route, but I have a more immediate problem with the backup taking 25 Hrs .
I also have two drives waiting to go in, was just hoping I could remove the offending one, and fit a new drive to remirror, then take a clean backup quickly, replace older drive, then the two new ones I can restore the good backup to.
I also have two drives waiting to go in, was just hoping I could remove the offending one, and fit a new drive to remirror, then take a clean backup quickly, replace older drive, then the two new ones I can restore the good backup to.
"Original IBM Firmeware same model and specs" replacement drive or "some spare" ?
ASKER
My question still remains unanswered. I need to know if what the Raid software is telling me is the truth. I cannot believe that a mirror set cannot be broken and a faulty drive replaced whilst the server is running. Someone must know this controller? If this is IBM then this has to be the dumbest thing they have done since losing out to microsoft!
Hot-swap backplane: just remove the faulty drive and insert a new IBM compatible larger or same size hdd. Rebuild should happen automatically.
Full backup to other hdd or tape first
Full backup to other hdd or tape first
ASKER
So you are 100% confident in removing the drive, even though the IBM server raid software insists the array will be unavailable?
I hope you are right, but it will have to wait for our next public holiday to prove it.
I hope you are right, but it will have to wait for our next public holiday to prove it.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
I had experienced really bad things with IBM Raid here (Exchange boot HDD fail, overwritting the healthy HDD in the mirror with crap. Officially fixed in the firmware one year later!).
The constantly rebuilding of the drives can kill both in the end , so it's time do something. I would go for a planned shutdown, maybe with an IBM guy on site equipped with a replacement HDD, or two (replace the second after the rebuild !). I know this won't make the customer happy, but a planned shutdown is what i would go for.