windows 2000 mirror set recovery

tcorrell · 08-15-2009, 04:29 PM

I'm running SBS 2000 on dual Pentium III Dell server with service pack 4. Exchange 2000 is running. No ISA.

I have a primary disk in a mirror set that is indicated as having an I/O problem in Disk Manager. (It has a bad block.) The mirror set contains two partitions: C:(system/boot partition) and D: (exchange data files). The two harddrives are scsi drives on an Adaptec 39160 controller. They are dynamic disks in Windows Disk Manager.

This primary disk has always required a boot floppy to boot. Never could figure out the problem with that. The primary disk though is still booting when the floppy is used. It is indicated to be healthy, online (errors) in Disk Manager.

I followed option 1 in Msft KB article 120227 http://support.microsoft.com/kb/120227: I tried to replace the primary disk with the error condition, by changing the SCSI ID to 0 on the mirror copy and installing my new disk and setting it to SCSI ID 1. The intention was to boot to the mirror copy and create a new mirror on the new drive. The scsi controller was registering the drives correctly. The mirror copy would not boot. It failed with a blue screen and stop error 0x0000007B inaccessible_boot_device as windows was about halfway loaded.

I did some online research and discovered Msft KB article 822052 http://support.microsoft.com/default.aspx/kb/822052. I put my primary disk with the I/O failure back into the machine and returned the mirror copy to it's original scsi id of 1. The machine booted to the primary disk (using a boot floppy). In Disk Manager, I found both disks now had the yellow warning. So I tried to reactivate the mirror copy assuming this was a mirror syncronization issue. The resync failed with error message VxSVC alert reporting a disk error on disk 0 and the sync stopped. The event log indicated that drive 0 had a bad block. It appears that the bad block stopped the resync effort on drive C:\ but Drive D:\ completed the resync.

I tried to use the boot floppy to boot to the mirror by changing the boot.ini file. (I changed the multi entry such that rdisk was 1 instead of 0). This failed with error message: can't locate ntoskrnl.exe.

So, I still have a booting server, but no fault tolerance even though Disk Manager is not telling me this. It just says Healthy(at risk) now on the C:\ drive. Is there any way to replace the drive with the I/O problem and return to a fault tolerant mirror set without reinstalling SBS?

Terry

**bilbus** · 08-16-2009, 08:27 PM

Why in gods name would you use software raid ....

Bad blocks are a comon issue with software raid, causes the mirrors to loose sync .. and wont re sync.

If you u pop out the disk with the issue, replace it with a brand new one you should be fine .. just kill the mirror and make a new one with the new disk.

08-15-2009, 04:29 PM	#1 (permalink)
tcorrell Registered User Join Date: Aug 2009 Posts: 1 OS: Windows 2000 server	windows 2000 mirror set recovery I'm running SBS 2000 on dual Pentium III Dell server with service pack 4. Exchange 2000 is running. No ISA. I have a primary disk in a mirror set that is indicated as having an I/O problem in Disk Manager. (It has a bad block.) The mirror set contains two partitions: C:(system/boot partition) and D: (exchange data files). The two harddrives are scsi drives on an Adaptec 39160 controller. They are dynamic disks in Windows Disk Manager. This primary disk has always required a boot floppy to boot. Never could figure out the problem with that. The primary disk though is still booting when the floppy is used. It is indicated to be healthy, online (errors) in Disk Manager. I followed option 1 in Msft KB article 120227 http://support.microsoft.com/kb/120227: I tried to replace the primary disk with the error condition, by changing the SCSI ID to 0 on the mirror copy and installing my new disk and setting it to SCSI ID 1. The intention was to boot to the mirror copy and create a new mirror on the new drive. The scsi controller was registering the drives correctly. The mirror copy would not boot. It failed with a blue screen and stop error 0x0000007B inaccessible_boot_device as windows was about halfway loaded. I did some online research and discovered Msft KB article 822052 http://support.microsoft.com/default.aspx/kb/822052. I put my primary disk with the I/O failure back into the machine and returned the mirror copy to it's original scsi id of 1. The machine booted to the primary disk (using a boot floppy). In Disk Manager, I found both disks now had the yellow warning. So I tried to reactivate the mirror copy assuming this was a mirror syncronization issue. The resync failed with error message VxSVC alert reporting a disk error on disk 0 and the sync stopped. The event log indicated that drive 0 had a bad block. It appears that the bad block stopped the resync effort on drive C:\ but Drive D:\ completed the resync. I tried to use the boot floppy to boot to the mirror by changing the boot.ini file. (I changed the multi entry such that rdisk was 1 instead of 0). This failed with error message: can't locate ntoskrnl.exe. So, I still have a booting server, but no fault tolerance even though Disk Manager is not telling me this. It just says Healthy(at risk) now on the C:\ drive. Is there any way to replace the drive with the I/O problem and return to a fault tolerant mirror set without reinstalling SBS? Terry

08-16-2009, 08:27 PM	#2 (permalink)
bilbus TSF Enthusiast Join Date: Aug 2006 Posts: 949 OS: OS2 Warp	Re: windows 2000 mirror set recovery Why in gods name would you use software raid .... Bad blocks are a comon issue with software raid, causes the mirrors to loose sync .. and wont re sync. If you u pop out the disk with the issue, replace it with a brand new one you should be fine .. just kill the mirror and make a new one with the new disk.