Link to home
Start Free TrialLog in
Avatar of Tito_Mahawk
Tito_Mahawk

asked on

Raid Controller fails, replacement procedure.

Hi experts,
Recently our server (IBM xSeries 220) began to had performance problems. Users complaining about problem accessing the files on it. At first it looked like one of the hdd was starting to fail (3 hdd Raid 5). I ran many test and they all went sucessful until i came to the Raid controller test (with ServeRaid, Boot then F2). Got something like 035-260-399-20070804-11-Raid Interface: Failed controller in slot 3.

So now, assuming the problem is my ServeRaid 4L Ultra160 SCSI Raid controller (right?) i'm planning to replace it (not a bad start?). SO the question is short and simple: HOW? (meaning without losing data)

I read i have better chance by using the same model (wich i found i can order, even if it's that old).
The perfect answer would be like:

1. DO this
2. Then do this
3. Then...
But i won't ask for too much :)
Thanks for helping
Avatar of SysExpert
SysExpert
Flag of Israel image

1) backup all data !!  Hard drives are cheap and you do not want to lose the data.

2) Call IBM , and see if the COnfig of the RAID is stored on the controller, on the drives or on both !

3) Write down any config info you can, and/  or export to a file, especially info availabel via the RAID controller

Check the IBM site for a Maint. manual for this controller or computer.

I hope this helps !
Avatar of Tito_Mahawk
Tito_Mahawk

ASKER

I have done mostly all of that. Backing up almost everyday (no image...). Waiting for IBM to answer my questions (i knew you guys we're gonna be faster). Have a DumpLog of all Raid and controller settings. HAve a pdf manual about ServeRaid 4L products and its says the raid info is stored on both controller and hdd. I guess i have all the info...just scared...
WHat i would hate to do is reinstall Windows, domain controller, policies...I did'nt build that system...
If you have some $$$, look at
Acronis Disk director to do an image.
tito_mahawk,
   It may be obvious, but I thought I would mention it.  Shoot for a day (like Friday at many companies) to attempt the replacement where there will be a good chunk (pref. 8 or 10 hours) of time where you can take down the system to do the swap.  You never want to do it under a time crunch... personal experience here... time crunch on these operations cause the horrible effect of jitters and you need your head on straight here.  Details are very important.  You also want to give yourself time to do some testing after the fact (diags, stress, client i/o, etc.)

Also, I usually suggest you go on ebay and pick up two of the same card.  Thow it into a test system and simulate your process.  ebay has tons of raid controllers cheap.  I use dell primarily so I keep two perc3/dc cards on hand for just such things.  I think I spent $120 total for both.  Pretty cheap when you rear component is on the line.

Best of luck to you.  You'll get though it.  :-)
-Cheers, Peter.
Oh one more point I just thought of...
  in the dell world, I have had the misfortune of improperly described actions programmed into firmware.  E.g.  The perc 3/dc controlers also store the container information on both the controller and on disk.  But the older perc 3/dc firmware would require you setup the containers again on the controller in the same manner as they were before.  After doing so it give you a friendly reminder to initialize the continer before you can use it.  That message is only meant when the container is new... and doing that initialization would wipe out your data...

Again, just reasons to be careful and practice before hand.
You might be able to ask IBM to walk you though the process...

-Cheers, Peter.
ASKER CERTIFIED SOLUTION
Avatar of Tito_Mahawk
Tito_Mahawk

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial