Jump to content

Degraded RAID5 Array

jaclaz
 Share

Recommended Posts

jaclaz

jaclaz

Posted
Posted

I am about to try re-setting up a HP Net Server after a move of a company office (actually the company was closed and I am supposed to keep the DATA archive).

It is an oldish Netserver LH3000 from HP, OS Windows 2K Server.

When re-switching it on after the move, one of the three 18.2 GB SCSI disks of the RAID 5 array on the machine had a red light.

Sure enough, checking it from the BIOS RAID utlilty gives me that the array is degraded.

I managed to talk to the ex IT guy (actually a girl) of the closed company and she said: "Sure, don't worry, it wasn't the move, it has been like that for a couple of years, never found the time to change the faulty HD". :w00t:

I backed up everything attaching another SCSI drive to the "other" controller, an Adaptec which has "normal" 68 pin connectors, just to be on the safe side.

Then I got a couple used 80 pin SCA drives from e-bay.

Now the question is:

Which is the "right" procedure to rebuild such a Raid array?

Connecting an additional drive setting it as "hot-spare" and let the machine do everything by itself?

Or manually starting a re-build?

If the latter is the right thing to do, anyone familiar with the Symbios controller that could spare some advice on how to do it, or any pointer to specific docs/procedures?

Thanks in advance for any advice/link.

jaclaz

Link to comment
Share on other sites

Tripredacus

Tripredacus

Posted
Posted

Is it possible to add more drives to an existing RAID5? Also consider the possibility of the controller/backplane being faulty and not so much the drives. On my old server (what 7Server used to be) it had a 3 drive RAID5 and was degraded. It showed 1 drive had failed, but after we replaced the faulty drive, the entire array broke and we lost all the data. It was only later when we couldn't create a new array sucessfully that it was determined that more than 1 port on the SAS backplane had failed.

So I recommend that you back up all the data you can before trying to do your repairs.

Link to comment
Share on other sites
jaclaz

jaclaz

Posted
  • Author
Posted (edited)
So I recommend that you back up all the data you can before trying to do your repairs.

You must be joking, right? :unsure:

I backed up everything attaching another SCSI drive to the "other" controller, an Adaptec which has "normal" 68 pin connectors, just to be on the safe side.

:whistle:

jaclaz

Edited by jaclaz
Link to comment
Share on other sites
ripken204

ripken204

Posted
Posted (edited)

lol, we all make those mistakes, but yes, backing up was a good thing to do. now you can play and not worry.

now what you need to find out is which hdd is down. replace that. then rebuild the array.

hopefully you have some sort of software for your controller that gives you the status of each drive.

OR

if you cant figure out which drive is bad..

since you already have it backed up, just test each drive individually, hopeful find the bad drive, check the status of the others while you are at it.

then rebuild a fresh raid5 array and copy the data back over.

Edited by ripken204
Link to comment
Share on other sites
jaclaz

jaclaz

Posted
  • Author
Posted
if you cant figure out which drive is bad..

I guess that the one with the blinking Red Light on it (and shown in SCSI BIOS as FAILED) should give me a hint. ;)

The question was a bit different, I have found reports on the net about different suggested procedures:

  1. hot-swapping the drive and let the **** thingy do it's business (through the Windows software)
  2. adding the spare drive in another tray with Server ON, set it as hot-spare (through the Windows software) and let the **** thingy do it's business (through the Windows software)
  3. hot-swapping the drive and start a manual rebuild (through the Windows software)
  4. adding the spare drive in another tray with Server ON, set it as hot-spare (through the Windows software) start a manual rebuild (through the Windows software)
  5. cold-swapping the drive and let the and let the **** thingy do it's business (through RAID BIOS extension)
  6. adding the spare drive in another tray with Server OFF, set it as hot-spare (through RAID BIOS extension) and let the **** thingy do it's business (through RAID BIOS extension)
  7. cold-swapping the drive and start a manual rebuild (through RAID BIOS extension)
  8. adding the spare drive in another tray with Server OFF, set it as hot-spare (through RAID BIOS extension) start a manual rebuild (through RAID BIOS extension)

Since I trust very, very little the Windows Software, not even knowing if anything in the (at the moment booting) system is gone beserk, I would instinctively go for #6 or #8, as they appear to me as the least "invasive" approach.

Or would it be better to go for #5 or #7?

Or the above does not make sense at all? .w00t:

:unsure:

jaclaz

Link to comment
Share on other sites
ripken204

ripken204

Posted
Posted

ya i would let the raid controller deal with it, not windows.

if the drive is dead then i see no reason in leaving it in there as it will be of no help to you, which is why i am for cold-swapping in this case.

i would have to vote for #7 since for #5, i am not sure how it will treat the new drive..

Link to comment
Share on other sites
CoffeeFiend

CoffeeFiend

Posted
Posted

But, have you backed up your data first? *ducks*

7 is your best bet IMO. Either ways, a lot of PERCs won't even give you a choice. When they see the new drive in there at boot time, they'll make you wait while they rebuild it, and only after that it'll let you boot an OS.

Link to comment
Share on other sites
jaclaz

jaclaz

Posted
  • Author
Posted

Just for the record and to finalize the present thread. :)

After some more research, I went for #6 (almost ;)):

6. adding the spare drive in another tray with Server OFF, set it as hot-spare (through RAID BIOS extension) and let the **** thingy do it's business (through RAID BIOS extension)

  • Switched Server ON.
  • Pressed Ctrl+M to enter SCSI RAID BIOS utility.
  • Scanned bus.
  • Failed drive was A0-01, working drives were A0-00 and A0-02
  • Added "new" disk drive.
  • Re-scanned bus.
  • Formatted newly added drive. (it does take some time)
  • Newly added drive gets "READY" status.
  • Set it as "HOTSPARE".
  • Let the thingy do it's magic. (it does take LOTS of time)
  • WAIT...
  • WAIT...
  • WAIT some more time...
  • During the process the drive is shown as "RE-BLD".
  • When the process is finished the new drive is shown as A0-01 and the old one is set as "READY" (though the latter is a lie :w00t:)
  • Check consistency (still some more time)
  • Status of Array changes from "DEGRADED" to "OPTIMAL".
  • Boot OS.
  • Everything works allright. :thumbup

All in all, a piece of cake, :yes:, but, as always better be safe than sorry. :whistle:

jaclaz

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share
Go to topic listing

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...