ICH9R Raid5 vanished, Offline Member disks

Turned my computer on the other day and my fears have finally hit me, my raid has gone :ouch:

I have a 6x500Gb Raid 5 running on my Gigabyte P35C-DS3R motherboard (which is a P35 Northbridge, ICH9R southbridge). Using F9 bios.

For some reason, when booting my computer today, instead of the usual green coloured list of hard drives in my raid, I've got a red list of "Offline Member" for each disk.

Also it says that RAID Volumes is "None Defined", yet my old raid use to be called "TheRaid"

So it basically seems that my RAID has vanished.

-------------------

What I want to do is:-

* Find out why this happened
* Try to recover important data
* Try to recover all data
* Maybe fix the raid so it works again

-------------------

What I have at my disposal:-

* The primary hard drive on the RAID machine is a separate drive which has never been a member of the raid. The OS is currently Windows Server 2008
* I may have access to a server at work where I can probably dump data (a few TB worth hopefully)
* I have various other PCs at home that I can use for net access or do other tasks.
* I have possibly another P35 chipset system at work I might be able to use.

-------------------

Previously this is what has happened on my system:-

I attempted to install Ubuntu 9.04 on the main OS hard drive (it was suppose to partition half of the drive so ubtuntu would be on it, while windows was on the rest). After checking this, it looks like it did. there are 3 partitions on the main drive.

Side note, I dont have any real programs on the primary hard drive as I only used this computer as a file server, and I probably threw on a benchmark or 2, but nothing of importance.

Anywho, I must have misread the info on Unbuntu's install but I'm pretty sure I all left it default, but after finishing the install of unbuntu, the RAID system (after BIOS post) was running in degraded mode, as the 6th hard drive had apparently been changed.

So I booted back into WinSvr2008 (note, I never got an option to choose between Windows or Ubuntu to boot from), and let the Intel Matrix Storage Manager rebuild the RAID over a 24hr period.

After that was done, I accessed a few files on the raid as normal but I never did anything overly exciting, then shut the computer down. WinSvr2008 never shut down, it just stays on the message screen saying "shutting down" so eventually I held down the power button to turn off the system.

So now a few days later, when powering up, the RAID screen after BIOS Post reports that "Raid Volumes: None Defined" and all 6 raid drives are sayign they are an "Offline Member".

Booted into WinSvr2008 and there doesnt appear to be a RAID listed in Intel Matrix Manager.

I've been checking forums for the past few days and people are reporting similar issues with "Offline Member" problems, but 'm not having much luck trying to work out their steps to fix it and how I can apply it to myself.

---------------------

Yes I have important data, and I'm sort of prepared to data recovery service.

Yes some of it is backed up on another machine, but the backups may be old, or are missing certain things. Plus I am currently rebuilding my backup-backup system so its missing more files than I'd like.

Suggest away people.

- Software I can use
- Phone numbers to ring Intel staff
- People to threaten
- Etc

Thanks :)
9 answers Last reply
More about ich9r raid5 vanished offline member disks
  1. Why did you install OS on your RAID5 array? Data on RAID5 array and OS/app should always be separate. Let it be a lesson taught.

    Quote:
    So I booted back into WinSvr2008 (note, I never got an option to choose between Windows or Ubuntu to boot from), and let the Intel Matrix Storage Manager rebuild the RAID over a 24hr period.
    Did you check to see under Intel Matrix Storage Manager software that rebuilding has actually completed before shutting off?

    AFAIK there's no chance of recovering data through software if the RAID5 array isn't recognised. It's been a while since I've last used Intel Matrix RAID. If there's an option to build the array, but NOT initialise it you could probably get your data back.
  2. Quote:
    * The primary hard drive on the RAID machine is a separate drive which has never been a member of the raid. The OS is currently Windows Server 2008


    As in my OS is NOT on the RAID

    The raid had finished rebuilding before the Offline Member problem started
  3. chiggenwingz said:
    Quote:
    * The primary hard drive on the RAID machine is a separate drive which has never been a member of the raid. The OS is currently Windows Server 2008


    As in my OS is NOT on the RAID

    The raid had finished rebuilding before the Offline Member problem started


    My apologies. I was a long-ish post and I was probably scan reading through it.

    As said in my previous post, you'll still have to somehow get the Intel Matrix RAID to recognise all those drives in a RAID5 first.

    I seem to remember something similar happening, not to me, but others. You can try this:
    1. Unplug power, clear CMOS by pulling battery out (just to be sure) and put the jumper into clearing position.
    2. Unplug all power connectors to the HDDs (including OS drive that's not part of RAID)
    3. After CMOS has been cleared you'll have to go into BIOS and retweak everything, remember to set SATA drive mode for ICH9-R to 'RAID' again.
    4. Power off then connect the first drive of the RAID5 array in (hopefully you'll see one 'online' green drive now)
    5. Do this one by one for the other 4 drives, leave the last one out
    6. Now connect your OS drive and boot into Windows, hopefully you should see the RAID5 array now in degraded, but readable state
    7. Plug the last drive of the RAID5 array in
  4. What would be the difference if I simply just cleared the CMOS and reset the system back to using the RAID with all 6 HDDs in my raid still plugged in?

    Why would I do one hard drive at a time?

    (I'm guessing I unplug the OS hdd to prevent it from booting into windows and automatically trying to do any modifications to the hard drives, like prematurley declaring it failed or rebuilding it?)
  5. Yep, unplugging the OS drive to avoid booting into Windows.

    Well, I suppose you could do it two drives at a time instead.
    The point here is to get every drive in the RAID array to go green/online during the detection phase.
  6. Sound and friend and I have found the cause of the problem, have a look at this other post here

    http://www.mail-archive.com/ubuntu-bugs@lists.ubuntu.com/msg1518800.html

    Now the next thing is to solve it. Few more things to work out and I'll post up details here when I'm done.
  7. Hmm you can recover the RAID5 on Linux/BSD. PM me if you want to follow this route. It will cost you some time and you probably have to allow network access so others can help you over the internet.

    But you can recover any RAID; no matter which system/driver made it. Because virtually all implementations store the data the same way; just not the configuration data (the meta-data). This data includes stripesize, disk member count, disk size, disk order, etc. If you know this manually, you can re-create the array using software RAID on Linux/BSD without writing to the drives. And you can try until you found the right combination and the partitions become visible and you can begin copying data.

    Again, this will take some time, so if you want my help make sure its worth it. But i can do it, assuming the data on your disks is intact and undamaged.

    BTW: did you use the "Write Caching" option in the Intel-drivers?
  8. Petty sure I disabled write caching.

    Gonna have Linux friend from work help me maybe, will fiddle with it at a later date.
  9. Since this thread has had over 10,000 views, which I'm guessing lots of people have had this problem.

    To follow up with this, I ended up having a friend do some magic in Linux, I have no idea what; but it enabled/initialized the disks or something? After booting into Linux I would then restart my computer (soft restart) and then select Windows to boot (as it was a dual boot system). THATS WHEN my raid array appeared correctly.

    I eventually decommissioned that computer mainly because I couldn't see the drive health and just never felt confident I could do a recovery reliably as a number of drives failed over time. Ended up just buying a NAS and using that instead.
Ask a new question

Read More

NAS / RAID Computer Hard Drives Storage