Out of the blue: 20min boot - system partition not accessible - RAID problem?

[Melanie77]

Hello experts,

I have a misbehaving server at our company.
The whole weekend with a solution. Any ideas on that?


Hardware:
P4, RAID1 with two 160GB Seagate Barracuda ST3160827AS, Raid Controller Promise Fasttrak 378 (Driver 1.0.01.37) on an Asusboard P4C800-E Deluxe

OS:
Windows small business server 2003

clients:
15 WinXP Pro clients

Partitions:
c: server's system partition
d: server's data partition

What happened around the time the server started to misbehave:
- Installed update for Lexware financial office pro (sql-based order mgmt system for our office) - installation worked fine without any errors
- Installed usual Windows update
- no hardware changes since server setup 4 months ago

Symtoms:
- cannot access system partition c: (neither via explorer nor via dir C: in dos shell), systems stalls and takes 5-10 mins to recover via TaskMgr
- cannot open control panel's "add/remove programs" - system stalls
- virus scan stops working when accessing c:
- takes 20mins to boot before win logo appears
- HOWEVER: server runs without _obvious_ problems once booted

Actions taken:
- tried each Raid harddisk individually - same behaviour each (takes ages to boot - runs once booted - cannot access C: or start add/remove programs)
- checkdisk finds no error - took ages to start though
- ontrack easy recovery pro v61007: harddisk check found no errors, but ontrack tool turned itself off when coming to partition c:
- Promise array mgmt tool (PAM 4.0.0.36) synchronize worked without error, array is said to be ok

My guess & questions:
- doesn't really look like a virus to me as booting takes ages before it even get to loading the OS and the server "works" once booted
- what can a software installation do to influence the booting process before windows is even started????
- no more ideas besides that...it's magic.

Thanks a lot
Melanie

[PCBruiser]

It sounds like possibly one of the Windows Updates may have really hosed something in the OS. Do you have a pre-update backup you can roll back to without losing much? You always could take the data off first and then do the rollback, then update the critical data files. My first thought would be a hard drive problem, but if both of the drives in a Raid 1 array behave the same, clearly that isn't likely. Given that behavior, something changed on both drives identically. I doubt the financial software update/upgrade caused the problem, so the only other possibility seems to be a failed Windowws Update. Possible??

If that isn't possible, do a full backup now, and try a recovery ("In-Place") reinstall post backup. If it was a Windows Update that caused the problem, that should back it out.

[PCBruiser]

BTW, I just noticed you just joined and this is your first post, so welcome to ABX, and enjoy your stay.

[Melanie77]

Well, if its a windows update problem it was the security update KB912919 which came up last week - I had to restart the server and the problem first occured at that time. My problem is that I cannot access the add/remove program feature - my first intention was to deinstall the last windows update and financial software update.
The backup/rollback is the very very last option I want to perform - I just checked the price on the asusboard and I think I will replace the hardware before rolling back the system -

by the way I know how to perform a rollback in XP with the >>%SystemRoot%\system32\restore\rstrui.exe<<, but where's the equivalent feature in that sbs2003????

[PCBruiser]

Good Q on the rollback, I personally use Acronis TrueImage and not the Windows rollback feature, so I haven't actually looked for it. Answer, I thought it was in the identical place for the server, but now that you ask, I don't think I even have it enabled since I am using Acronis.

[Melanie77]

something in addition:
we tried booting from CD (BartPE) - so no server OS involved - and it still took ages. Hit the switch - then the BIOS - then the Promise BIOS (finds array and says it's ok - and then there are 20 long minutes - so it must be some hardware problem....

[PCBruiser]

Now, that's interesting. I think you are on to something. Good thought.

[Miles]

Hi Melanie and Welcome to ABXZone

Now on to your issues, first let me say there are quite a few people here that know Windows 2003 Server far better than I ... I spend most of my time on Linux/Unix Servers at work, but a few things that come to mind are:

1. Have you checked the error logs?

2. Is the Windows Server 2003 Volume Shadow Copy Service enabled read here If so you can roll back, if not I don't think so.

3. I'm wondering if something hasn't slightly corrupted the MBR, not totally disabling a boot but causing your system to take an excessively long time to boot. Might want to check the C:\boot.ini file for corruption.

Don't worry the Windows 2003 Cavalry will arrive later and help much more than I can.

[Melanie77]

Aaaargh.

1. No errors in the error log.
2. No Shadow copy enabled - no rollback.
3. MBR-thats a hint

What happened so far:
1.
We digged out an image of a previous server installation of this server and plugged it in to check if the hardware has a problem. I was flawless. So obviously no hardware problem with the board...this would have been the easiest solution:....-(

2.
I put one of the RAID disks to an IDE controller - same problem with booting, accessing c: manually etc.

3.
Wanted to make a fixboot & fixmbr and wanted to backup the whole disk before that. The backup completely failed an left us with a whacked array....so now we're back at synchronizing the two RAID images which hopefully works.

New ideas:
1. Chkdsk doesn't find anything - but there must be some problem with either the partition, mbr or boot sector.

2. How can the server run at all with a damaged partition or faulty booting??????Because thats what it did all the time the last days.

Any ideas for really good tools to check on these?

[pointreyes]

Did you run the hard drive diagnostic on the drives? Not the Promise diagnostic but the hard drive manufacturer's utility. You can safely do that without the drives being on the RAID. Also, I believe you will need to look up how to recover SBS2003 with Active Directory-that makes a difference from what you usually do with XP Pro. Recovering an Domain controller in AD is not that much fun.

[Melanie77]

Thats next on the schedule - it seems the sync worked and the ******* is currently booting again......