Server 2003 rebooting on startup

[edie209]

I came back to work today after a week off to find one of the drives in a Dell Poweredge 2850 had failed, the system is set up with raid5 using 4 drives and a hot spare. I downed the server and replaced the drive, but on reboot the server reboots at the point that the message "Windows is starting up" is displayed (about 5 seconds into it).

 

Server 2003R2 sevice pack 2

 

This happens in safe mode too has anyone encountered this and know a solution? please

[Uraken]

try booting from a windows cd and go into recovery console and running chkdsk/r

[netadmin]

I had a similar issue, but never found out what caused it.

The solution that worked for me was to "Revert to last known good configuration."

 

It lost a few setting changes, but worked just fine.

[edie209]

Ok I rebuilt the drive overnight and all was complete when I got in this morning. so I rebooted the sever and got the message

c:\windows\system32\config\system is missing or currupt

 

So I rebooted and went into recovery console to run chkdsk /r as mention by Uraken on entering the command chkdsk /r I am presented with a message

cannot find autochk.exe  

 

So I decided to look on the c: drive for autochk.exe and and i find I am unable to use the dir cmd on the c drive and I get the message

An Error occurred during directory enumeration

 

So I am running out of options unless anyone here has a suggestion??

[Geoff]

Looks like a failing hard drive to me.

[tech_guy]

Have you tried chkdsk /p /r

?

[edie209]

I think I have a way of fixing the server in c:\windows\repair the file "system" is backed up, if I could find a way of accessing that file and then overwriting the the file in c:\windows\systems32\config I think I might be able to get the server back up but I can't do this through recovery console is there any other way that anyone knows.

 

Tech-guy I tried your idea thanks but I get the same problem cannot find autochk.exe

[Geoff]

A Linux boot cd will allow you to access the file system 'offline'. SystemRescueCd is my favorite for doing this:

 

http://www.sysresccd.org/Main_Page

[Butuz]

Sounds like your array could have become corrupt. Did your hot spare actually come online and succesfully do a consistency check?

 

Andrew

[edie209]

Geoff downloading now thanks hopefully that will work.

 

Butuz the hot spare seems to be ok

[edie209]

Well unfortuanatly that didn't work so I have had to reinstall the OS and then restore the system state this went ok but I have been left with quite a few errors now which I will look into tomorrow.

 

One question though has anyone restored the system state o a domain controller? because although I can access Active Directory Users and Computers I have to do it through Manage your computer there is now no shortcut in the menu along with other Active Directory tools.

 

I haven't sevice packed it yet but that shouldn't cause a problem should it?

[PiqueABoo]

Shouldn't you have restored c: drive + system state?

[Geoff]

Nah, I wouldn't even attempt it. Just bring the machine up from a fresh install and force a replication from your other DC(s). A normal dcpromo will be sufficient to trigger this.

 

If you're restoring a possibly broken backup onto the DC with possibly suspect hardware you risk introducing errors into your AD environment, which will propagate to your other DCs. The result is incredibly depressing and almost impossible to fix.

[maniac]

I agree with geoff there. Only restore your active directory from a backup if it's the last resort and you've completely lost you ad structure. If you have another functioning domain controller then best to do a clean install, fully update and patch it, then DCPROMO it as stated above.

 

If there were other files and shares on the server, these can be restored from a backup after all the above is complete, and should just work as they did before as the SIDS and file sharing information should be preserved on the backup.

[edie209]

read your posts last night so I remoted in and shut down the sever and I am now installing a fresh copy as per your suggestion and will let it replicate with the other domain controller.

 

Thanks for your contributions guys sometimes its just good to know your not on your own

[PiqueABoo]

Nah, I wouldn't even attempt it.

 

Chicken ;b

 

I'll conditionally surrender that point. The last few I've been near had things other then AD they needed (or worse, bits of AD they needed), but if it is mostly a "pure" DC then yes building it as new DC is less risky.

 

I'd check what was going on with replication on the working DCs first though. If they were happily replicating from the broken DC then fine, otherwise you will lose any significant changes made to AD on the broken DC when replication wasn't working.

 

And obviously there's some cleaning up to do afterwards (more if DC had the FSMOs).

[edie209]

Just to let you all know I have found the cause of the problem according to Dell the Poweredge 2850 MUST use the dell raid controller drivers NOT the Microsoft ones and also the scsii drives must be using the latest firmware. apparently the parity can be effected over wise, as in my situation