cannot book windows 2008 server, getting unmountable boot volume
Hi, I have a client who has a windows 2008 server that went off a few times over a powercut and never properly shut down. Due to the UPS being faulty. which we have also just discovered :(
It's a Dell poweredge T110 II with a SAS perc s100 intergrated Raid card running RAID 0 with 2 hard drives.
When I turn the server on, I get unmountable boot volume.
We have tried doing last known good, safe mode and the repair option.
If I boot off of the dell supplied windows 2008 installation disk, it doesn't find the drive, so I figured I may need to provide the drivers, I go to add the drivers, it then says adding, goes to refresh the display. I'm assuming to show the operating system. and it just hangs on the refresh, not showing any thing. I have left this 6 hours.
Please also not, when you run the windows repair off of the local install, it sits on trying to load the restore program and doesn't actually make it to the next step.
I'm lost on what the problem could be and need to get to the data, I've also learnt that no one has been keeping an eye on the backups and fear they may not have worked for a while, if at all.
The raid controller says the raid is ok. Could it be the raid card, raid itself or windows that's faulty ?
any ideas
It's a Dell poweredge T110 II with a SAS perc s100 intergrated Raid card running RAID 0 with 2 hard drives.
When I turn the server on, I get unmountable boot volume.
We have tried doing last known good, safe mode and the repair option.
If I boot off of the dell supplied windows 2008 installation disk, it doesn't find the drive, so I figured I may need to provide the drivers, I go to add the drivers, it then says adding, goes to refresh the display. I'm assuming to show the operating system. and it just hangs on the refresh, not showing any thing. I have left this 6 hours.
Please also not, when you run the windows repair off of the local install, it sits on trying to load the restore program and doesn't actually make it to the next step.
I'm lost on what the problem could be and need to get to the data, I've also learnt that no one has been keeping an eye on the backups and fear they may not have worked for a while, if at all.
The raid controller says the raid is ok. Could it be the raid card, raid itself or windows that's faulty ?
any ideas
ASKER
Hi Mark
It's a raid 0 setup (stripe) not mirror raid. If I disconnect the 2nd drive, then I will mess up the raid won't I.
The problem with the Windows disk idea, is there is no driver / hardware support for the perc s100 on normal operating systems. Servers only.
I can't tell you how big the partitions are, as I can't see them.
My problem is, I can see it loading and fail, but can't seem to see the partitions.
I'd be more than happy to get the data off and then rebuild the server as suggested. Just can't get to the data :(
It's a raid 0 setup (stripe) not mirror raid. If I disconnect the 2nd drive, then I will mess up the raid won't I.
The problem with the Windows disk idea, is there is no driver / hardware support for the perc s100 on normal operating systems. Servers only.
I can't tell you how big the partitions are, as I can't see them.
My problem is, I can see it loading and fail, but can't seem to see the partitions.
I'd be more than happy to get the data off and then rebuild the server as suggested. Just can't get to the data :(
Your quite right if it's a raid 0 and you remove (or fail) a disk you have lost all data..
You can build a PE boot environment with a server OS.. that works fine. you can also include the driver for the s100 in the build.
You may find later versions of Windows Server eg. Win2k8r2 or Win2k12r2 have the required driver already integrated into the OS DVD.
Can I suggest never building a server with a raid 0 array..we'll not unless it's a LAB server.
If it's important, make it somewhat robust..
Over the weekend I can see we have had 2 disk failures.. fortunately on different servers.
You can build a PE boot environment with a server OS.. that works fine. you can also include the driver for the s100 in the build.
You may find later versions of Windows Server eg. Win2k8r2 or Win2k12r2 have the required driver already integrated into the OS DVD.
Can I suggest never building a server with a raid 0 array..we'll not unless it's a LAB server.
If it's important, make it somewhat robust..
Over the weekend I can see we have had 2 disk failures.. fortunately on different servers.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
i agree, I never do raid 0, raid 1 and upwards for me, there's no fail over with raid 0
It was already like this before we took over.
It was already like this before we took over.
The first thing I would do is get a disk (usually disk 1or the 1st disk in the raid set).
I really need a Dell expert to advise there but I believe the percs100 is a hardware raid controller and therefore you should be OK to:-
power off the system,
remove disk 1
power back on
After the raid controllers bios reports a disk is off line, you choose OK to continue loading.
At this point you don't care about if the system boots of not.. You now have a disk in your hand what is you point of recovery assuming everything else does not work.
Put a new disk (of the same size) into disk bay1 and let the raid controller mirror to disks again.. (disk2 willlbe seen as the master as disk 1 is known as failed and recovering.)
Please remember why I'm suggesting this.. I'm at this point not trying to get your server bootable. I'm only getting you to ensure you have a copy of the original system.
In theory you will be able to attach this disk to a SAS controller which is installed in a running PC or Servers (maybe via a breakout cable) and read the file system. which means you can get data back, but lets not go there quite yet..
Back to recovering the actual server.
As you've mentioned.. there may be hardware damage due to the power outages..
Ideally I'd suggest you take the disks and put them into another server test if the hardware is at fault.. or remove the 2 mirrored disk, clear the server configuration, install 2 new disks and build a fresh OS on it again to prove there is no hardware issues.
But this routine is dependent on you having some disks available to you.
Can you confirm what the partition layout is..
eg.
1. SystemReserved 100MB boot loader
2. Operating System C:\ 100GB+ OS Install location.
3. Data volume D:\ ?? GB
Thinking about this.
The server boots or at least finds the boot loader..Assuming that is a small 100MB partition which only has the loader installed in it.
I think at this point I'd be booting off a Win7 or Win8 PE CD-ROM and running a check disk of the core file system.