[Last Call] Learn how to a build a cloud-first strategyRegister Now

x
?
Solved

RAID broken after BIOS update - how do I re-do it?

Posted on 2010-01-06
17
Medium Priority
?
509 Views
Last Modified: 2012-05-08
Hi
Prior to Xmas I tried upgrading the BIOS on my server (WIN 2003 SBS) to help some errors with a SCSI tape drive and I have now broken the RAID array which was mirrored across 2 hard drives. I'm desperate to get things back to the way they were.

The software I am using is the Intel Matrix Storage Manger that is compatible with the chipset on the motherboard for RAID array. The motherboard is a Gigabyte P45 ICH10R.

I have since found out that in some cases the computer will blue screen and re-boot after trying to set the RAID again in the BIOS if you already have the OS installed. This happened to me so in desperation I put the BIOS back to the old one and set it to normal volumes and no ACHI or RAID. This was the only way it would boot up again and now I have 4 separate volumes instead of 2 mirrored. Help!
The Intel help on the Storage Matrix software does say you can make a RAID after the OS is installed but the necessary prequisites don't seem to work for me.

The Matrix Storage software now has a error when starting up and I cannot use my imaging software anymore either. I tried making the volumes dynamic in windows but that has not helped and of course it won't allow me to mirror as it thinks I have no spare space/drive which I don't .

Can anyone please help me to get the current volumes back to a mirror? Thanks
0
Comment
Question by:TracyFazackerley
  • 9
  • 8
17 Comments
 
LVL 47

Expert Comment

by:David
ID: 26197395
I would check release notes of the BIOS.  It is possible that the upgrade modified the metadata, and the downgrade munged it.  You may only be able to get it fixed by using a 3rd party package like runtime.org's reconstructor to recover the entire array, then rebuild the RAID from scratch with new firmware, then restore.
0
 

Author Comment

by:TracyFazackerley
ID: 26197482
At the moment I have 4 volumes (C and H) and (D and G) which I'm not sure where all the new data is going on the server as we have still been using it for a couple of weeks with no problem accessing all the folders/files. I guess it just points us to local disk C and Data volume D. So 2 volumes will be different now?

Do I need to recover data using the 3rd party when I can still see it?
0
 
LVL 47

Expert Comment

by:David
ID: 26197683
Sorry, you didn't say the volumes were good or bad, just that you had 2 sets of them.  So you basically instead of RAID1, you have a JBOD config.  Does the BIOS report this is a JBOD confiig, or a RAID1?  IF bios says it is JBOD, then you should be able to tell it to convert the disks into a RAID1.  In this case it is probably best to contact intel support. It may be nothing more than a few keystrokes to "upgrade" the JBOD into a RAID 1 so it works.

0
Transaction-level recovery for Oracle database

Veeam Explore for Oracle delivers low RTOs and RPOs with agentless transaction log backup and transaction-level recovery of Oracle databases. You can restore the database to a precise point in time, even to a specific transaction.

 

Author Comment

by:TracyFazackerley
ID: 26205528
Hi

The volumes are still good, just lost the RAID 1 with trying to upgrade BIOS. When I did set the BIOS to RAID it would not boot up and it kept re-booting. That is why I left it in normal config in the BIOS. Maybe there is another setting in BIOS I have missed as the BIOS upgrade set the BIOS to defaults.

I will look at it again.
0
 
LVL 47

Expert Comment

by:David
ID: 26205641
You would think the release notes would have addressed such things.  (Actually they usually do, did you check? I know I am guilty as the rest of us for not reading release notes, but I DO make it a point to read them if I am doing anything that can affect a RAID controller).   Moving forward, I think it will  let you convert a non-RAID disk into a RAID1 via a windows-based utility you run, but since the RAID config got zapped with an upgrade then there may be a mismatch between the metadata in the NVRAM on the motherboard and what it puts on disk drives.

As this is also a win2k3 SBS machine, which is undoubtedly vital, then you need to be conservative.  I would do a backup first (one that lets you do a bare metal recovery) to a scratch disk, then start fresh with a new RAID1 config, format it on RAID controller to clear out any bad blocks, then do the restore.

This is more time consuming but safer .. just make sure that if you put all the eggs in the basket of a single scratch disk for backup, that you have 2 copies, or that you backed it up to a RAID-protected array on the network. Even backup disks die.
0
 
LVL 47

Expert Comment

by:David
ID: 26205700
P.S. I know  you are tempted to just skip the backup step, and IMHO you have really great odds that just telling the RAID to turn the JBOD disk into a RAID 1.  If this was my company's SBS machine, I would just do the right thing and schedule it for weekend and do backup first.   I've had to rebuild my own SBS 2003 system before and had to spend well over $1000 to some consultant to get my exchange going properly because the restore didn't work properly.  It just isn't worth the headache of not backing up before changing RAID configs.

Don't forget to make 100% sure you are running latest RAID firmware and drivers, and read release notes before doing this. Nobody likes surprises.
0
 

Author Comment

by:TracyFazackerley
ID: 26205826
It all sounds long and tedious but If that's the only way, then it will have to be done. I will check first the Inel support site again and try once more in the weekend to change the BIOS setting to RAID.
It may be that there wasn't an update even as this motheboard is only a year old. I might have put the same one on anyway to start with but still setting all BIOS settings to default obviously didn't help.
I wasn't 100% certain I was looking at the right numbers for the BIOS version.

I will post how it all goes after the weekend.
0
 
LVL 47

Expert Comment

by:David
ID: 26205911
well, here is a big part of equation.  You need to figure out if the RAID controller reserves any blocks of the HD for metadata, and if when you have it in JBOD, does it reserve same number of blocks.

So ..
1. Look at disk drive mfg web site and get total number of blocks, not MB, but exact number of blocks.
2. look at the device manager disk properties.  See if the block count is the same. (physical disk, not partition, of course).

If the windows O/S sees exact number of physical blocks as the specs report, then you know that the controller is not reserving metadata.  For your particular controller, the question for intel is, "When I put controller in RAID mode, does it reserve any physical blocks for Metadata?"  If answer is yes, and your current config is using 100% of disks, then you know you currently have no metadata, so the usable blocks in RAID1 mode is less, which means that migration is risky as it will have to change partition table.  This is scary and shouldn't be attempted w/o full backup.

If a few MB of data is missing, then this is good indication the controller already grabbed your metadata, so migration should be painless (but still back up, it is best practice and you need to CYA)
0
 

Author Comment

by:TracyFazackerley
ID: 26206294
On Seagate website could only find the model ST3500418AS has:

976,773,168 sectors
16,383 cylinders

Do either relate to blocks?
0
 

Author Comment

by:TracyFazackerley
ID: 26206358
System Information on the server tells me the sectors are slightly less at 976,768,065 so does that mean the controller has the metadata?
0
 
LVL 47

Accepted Solution

by:
David earned 2000 total points
ID: 26206390
Yes, this leaves exactly the right amount for the metadata, so the resulting partitioning on a RAID1 should remain unchanged, and you should be able to migrate in-place from JBOD->RAID1 by just selecting the original drive as the source, and the other drive as the one you want to be the target of the cloned data.
0
 

Author Comment

by:TracyFazackerley
ID: 26206448
Well some good news then. So will I have to format the 2nd disk first so the RAID will migrate. It has most of cloned data on at present, same as 1st disk.
If I do that first will it stop the blue screen/re-boot when I change the BIOS to RAID config?
0
 
LVL 47

Expert Comment

by:David
ID: 26206468
Yes, format it just to clear out the metadata at the tail end before attempting a migrate.  This also is a nice write diagnostic that will automatically map any bad blocks, so your RAID controller doesn't have to later on.  (And, I will bug you one more time, do a backup )
0
 

Author Comment

by:TracyFazackerley
ID: 26206523
Ok thanks I will definitely do the back up, we have an external large drive I can do this to as my imaging software won't work until I get this sorted. I can use Win Backup.

Hopefully I can boot ok when I set up the BIOS and then all should be well.

I'll let you know how it goes after this weekend. :)
0
 

Author Comment

by:TracyFazackerley
ID: 26240351
Update:

I have successfully upgraded the BIOS (it was 2 versions behind it seems) and turned on RAID in the BIOS without the server re-booting afterwards. This may have been because I deleted the 2 extra volumes on the 2nd disk prior to turning on RAID?

I also upgraded the Intel Storage matrix technology to latest 9.5 version. I can see the disks ok in the Intel software but it will not give me the menu to create an array volume. disk 0 is C: (boot) and D: (system) while disk 1, I just made volume F: They are both same size.
If I put in another internal disk it will allow me to create a mirror volume between disk 1 and the 3rd disk ( I didn't do this as much smaller capacity on 3rd disk)

So how do I get it to migrate a mirror from system disk with OS to the 2nd disk? Do I have to have 2 volumes on the second disk exactly same as system disk? I thought it just created these. Or should it be unallocated disk?
0
 

Author Comment

by:TracyFazackerley
ID: 26460344
Figured it all out with a bit of trial and error but have the RAID back again thanks!
0
 
LVL 47

Expert Comment

by:David
ID: 26460464
You are welcome, make sure you run regular consistency checks on the RAID. This prevents data loss if you have a drive failure AND a bad block on one of the surviving disks.
0

Featured Post

Free Tool: Site Down Detector

Helpful to verify reports of your own downtime, or to double check a downed website you are trying to access.

One of a set of tools we are providing to everyone as a way of saying thank you for being a part of the community.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Steps to fix “Unable to mount database. (hr=0x80004005, ec=1108)”.
Exchange administrators are always vigilant about Exchange crashes and disasters that are possible any time. It is quite essential to identify the symptoms of a possible Exchange issue and be prepared with a proper recovery plan. There are multiple…
In this Micro Tutorial viewers will learn how to use Boot Corrector from Paragon Rescue Kit Free to identify and fix the boot problems of Windows 7/8/2012R2 etc. As an example is used Windows 2012R2 which lost its active partition flag (often happen…
This tutorial will walk an individual through locating and launching the BEUtility application to properly change the service account username and\or password in situation where it may be necessary or where the password has been inadvertently change…

834 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question