Solved

Dell PowerEdge 2950 Raid 5 Drive in Foreign State will not boot Server 2003 OS

Posted on 2009-05-20
4
12,244 Views
Last Modified: 2012-05-07
There was a previous question asked by someone stating the following and I am having the same exact problem:
-------start---------
Today morning we've found an server that has somehow crashed over the weekend.

It's an Dell Poweredge 2950 with an Perc5/i adapter and 2 attached 70 GB SAS drives  for raid1,  and 4x 130gb SAS for Raid5.

If we try to boot the server the Perc5 adapter shows an error:
"Foreing configuration found. Press "c" for config utility, any other key to continue..."

"Continue" will result in try to boot from PXE-dhcp and repeat. Win2003 will not boot.
--------end---------
The difference in my issue is I have a 4 disk array and one of the disks is showing up in a foreign state.  The resolution to that particular issue for that guy was he deleted the virutal disks and recreated them and everything fixed itself... I am afraid to try that without at least posing the question, will that really work?  The drives appear to not be corrupted in any shape form or fashion when I go into the configuration manager - it allows me to "blink" all the drives so connectivity is good and it's reporting the drive sizes correctly and other than the "foreign" state of the one drive, everything else appears to be ok?

Any ideas on how to proceed from here?

Thanks in advance!
0
Comment
Question by:VaiSatchAtrucci
  • 2
4 Comments
 
LVL 4

Accepted Solution

by:
gardmanIT earned 500 total points
Comment Utility
Just to avoid confusion / errors.

I am assuming here that the RAID 1 pair have your OS installed and the RAID5 set (4 disks) is used for data and is not involved in booting the server. I also infer from the question that the "foreign" disk is one of the 4 RAID5 disks.

This would be the correct way to have the server set up as you do not want raid 5 involved in the boot process as if the RAID card loses its memory raid 5 disks can't be "recovered" to boot from whereas eith of your Raid1 mirroered pair should be bootable without any Raid data from the card.

Problem.
For some reason your raid card seems to think one of the disks in the raid 5 set is a "foreign" disk, this means the disk has probably failed and is not returning the serial number it did before. As a result the raid card see's the disk as foreign (it thinks its a disk its never seen before and this was un-expected to it),

Best Solutiion.
There should be no need to destroy the Raid5 set at all. Raid 5 is redundant so you can afford to lose 1 disk in a raid 5 set without losing the data or access to it (the machine will access this drive more slowly when a disk is failed). You need to carefully identify the disk shown as "foreign" in the raid card config, get the config utility to flash the lights on the disk, SAS disks on the Poweredge 2950 are hot swappable so go ahead now and pull this disk out of the unit.
Your raid card will now see something it can handle, a failed disk in a Raid 5 set. You will get some amber lights and warnings that the raid 5 set is degraded but your server should now boot.

Now to resolve the "foreign disk" This disk may be faulty - hence why it was seen as foreign, or it may have had a glitch. If now its booted you shove it back in the raid controller will either, see the original disk or see a new disk and rebuild the "missing" data on it from the other 3, or see it as faulty and ignore it.
Sorry I can't remember if the RAID card on the 2950 does this automatically, I dont think it does so you can use Dell Server Administrator in windows to check the status of the disk once re-instralled and start the rebuild process.
If the disk shows as faulty order a replacement PDQ as you have no redundancy left on the raid 5 set until the disk is replaced and it will be slow performance from the set till its replaced and the rebuild is complete.

Good Luck,

Ian.

0
 
LVL 2

Expert Comment

by:Hoard
Comment Utility
The system boots now currently right?
If so; either use Openmanage Server Administrator to clear forgeign or go into the CTRL R for the RAID as the system is coming up.
CTRL R will probably be the easiest.
Hit that as the system is coming up.
Once in there; on the first page; highlight the controller (It'll be the top option in the tree)
Hit F2. You have an option to Clear Foreign.
Say yes.
That'll take care of that problem.
At that point I'd recommend setting this problem drive as a hotspare and see if it will rebuild. If it does; you're good to go (Update RAID firmware after the rebuild is finished)
If it doesn't you're problly looking at a bad disk.
Good luck!
0
 

Expert Comment

by:GreyWolfComputers
Comment Utility
I have the same problem that seems to have started when the owner removed one of the drives and it showed up as foreign. During the eval process, another drive was removed and it too came up as foreign. I tried the "clearing" outlined above. That cleared the "foreign" status but now they show up as "missing". I'm reluctant to rebuild in the event I might lose critical data. Any suggestions.
0
 
LVL 4

Expert Comment

by:gardmanIT
Comment Utility
Hi Grey Wolf,

If you have two "missing" drives in a raid 5 set you have lost the data and a rebuild will not work.

RAID 5 utilisies disks as follows and allows for 1 disk in the set to fail without loss of data.
you can have any number of disks from 3 upwards (there is often an upper limit due to controller capacity or slots) These disks are used by the RAID controller as follows. Each block of data is split across (using a 3 disk example) 1/2 the data on one disk, the other half on the next and whats known as parity data on the thrid. In a four disk raid set its 1/3 of the data on the first 3 disks and parity on the fourth. This is why in RAID 5 you get the capacity of the number of disks -1 as one disks worth of space is used to store parity info.
In effect the parity info is a calculation of a checksum, a mathmatic sum of all the bytes of data.
If you lose a disk it will either contain the parity info or 1/2 of the data, if the parity info is lost no problem the data is still on the other two disks, if half the data is missing the RAI controller can take half the data and the parity info and do the maths to reconstruct the missing half of the data.

So RAID 5 lets you lose a sinlgle didk and recover or "rebuild" it onto a replacement disk, if while one disk is down however another fails (or was removed in your case) you have lost the data. Rebuilding won't work as too much of the data / parity is missing.

Trying a rebuild will not do any harm but it will probably just confirm that you have lost the data when it fails.
If you have a disk fail in RAID 5 the normal process is....
Remove it (having made sure you get the right failed drive - blink the drive if possible from the RAID controller or its software (say dell open manage)
Insert a new drive - it wil show as foreign which simply means the controller sees it as a disk it hasn't seen before, clear the foreign status.
Select the virtual disk (the whole raid 5 set) and select rebuild - this intitiates the process of calculating all the missing data and or parity data that was stored on the failed disk and writes it to the new one.

For info, if the RAID controller had had access to a spare disk already installed in the server as a "hot spare" when the first disk was removed it would have automatically rebuilt the missing data on this spare so when the second disk was removed it would not have been ternminal. Another option is RAID6 (not supported on the Perc 4 in the 2950 from memory. This uses an extra disk in the RAID set for parity info so in effect needs 4 disks min but allows two disks to fail (be removed) without failing. Down side is your capacity drops to number of disks -2

0

Featured Post

What Security Threats Are You Missing?

Enhance your security with threat intelligence from the web. Get trending threat insights on hackers, exploits, and suspicious IP addresses delivered to your inbox with our free Cyber Daily.

Join & Write a Comment

My purpose is to describe the basic concepts of virtual memory as implemented in a modern Windows-based operating system. I will also describe the problems inherent in older systems and how virtual memory solves them. The dark ages - before virtu…
In a hurry?.. scroll down to "HERE's HOW TO DO IT" Section. Greetings All, I was going to post this as question/solution, but its seems more appropriate as an article considering its length.  I felt it important to illucidate all the details c…
This video Micro Tutorial explains how to clone a hard drive using a commercial software product for Windows systems called Casper from Future Systems Solutions (FSS). Cloning makes an exact, complete copy of one hard disk drive (HDD) onto another d…
With the advent of Windows 10, Microsoft is pushing a Get Windows 10 icon into the notification area (system tray) of qualifying computers. There are many reasons for wanting to remove this icon. This two-part Experts Exchange video Micro Tutorial s…

743 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

Need Help in Real-Time?

Connect with top rated Experts

16 Experts available now in Live!

Get 1:1 Help Now