C0000218 registry file failure Registry cannot load hive (File): \SystemRoot\System32\Config\Software problem and tactics for ease.

jazzIIIlove
jazzIIIlove used Ask the Experts™
on
Hi there;

I have Windows 2003 R2 Server having a VM of another 2003 using VMWare Server in it.
I have the error and the system "was" not booting even in safe mode after I was trying to have a snapshot of 20 G VM disk:

C0000218 registry file failure Registry cannot load hive (File): \SystemRoot\System32\Config\Software problem and tactics for ease.

Since I had XP CD, I boot and use recovery console to checkdisk and it didn't fail luckily, but it finds lots of bad sectors in D: drive having a problematic VM of VMWare Server.

Now, after the checkdisk, the system boots normally. My questions are:
1) Could a corrupted VM cause the physical machine to have bad sectors?
2) How can I heal the VM? (No snapshots, no system restore (in fact 2003 has no system restore) Is it possible to heal it?
3) If I delete or transfer that VM, could those bad sector be gone/healed?

Best regards.
Comment
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
Commented:
The VM uses the hard disk in your computer that it is assigned to so if bad sectors are found then that means their were bad sectors on your real hard drive inside your server unless you did something like moved the VM or anything else that could corrupt the drive.

you can perform a CHKDSK and a Repair on a server OS but the repair is not recommended and can screw up AD and Exchange records so it is not recommended. Its recommended to have some sort of data redundency such as a RAID in place of relying on a Windows Repair or Restore.

You should never have to "System Restore" a server
Hi;

>>unless you did something like moved the VM or anything else that could corrupt the drive.
Exactly, the first time I want to move the VM, by simple copy-paste, it gives CRC error, then right now, it turns to redundancy error. Prior moving the VM, everything was ok.

>> the repair is not recommended and can screw up AD and Exchange records so it is not
>>recommended.
There is no AD or Exchange but Oracle 10g in both physical and VM.

>>Its recommended to have some sort of data redundency such as a RAID in place of relying on a >>Windows Repair or Restore.
There is no RAID as far as I know in this server. I will check it.

>>You should never have to "System Restore" a server
Why did you say so? If I have a restore point, I can go back?

Best regards.

Commented:
Thats not how Server OS's work.
System Restore is for home users that dont know any better.
Acronis in Gartner 2019 MQ for datacenter backup

It is an honor to be featured in Gartner 2019 Magic Quadrant for Datacenter Backup and Recovery Solutions. Gartner’s MQ sets a high standard and earning a place on their grid is a great affirmation that Acronis is delivering on our mission to protect all data, apps, and systems.

Then how can a serrver administrator recover or restore if there is a problem? I mean I don't find this sensible since you may need to have some kind of system restore feature built-in. Why does an admin want to pay a 3rd party software since there should be already one in his/her 2003?

For my case, what should I do? (1, 2, 3)

Best regards.
Commented:
1 - Reboot the server (the real server not the VM) use the windows disc to get into the command prompt / recovery console and run CHKDSK /R
2- VM Snapshots is the only way to go back on a VM, if its not a VM you need a backup program like Acronis
3- The bad sectors will transfer with the VM until you run CHKDSK /R to remove them from within the VM
>>1 - Reboot the server (the real server not the VM) use the windows disc to get into the command >>prompt / recovery console and run CHKDSK /R
That I had done as in my original post successfully.

>>3- The bad sectors will transfer with the VM until you run CHKDSK /R to remove them from within the >>VM
I see, so I will implement chkdsk /r to the VM also.

Best regards.

Commented:
Good Luck :)
Top Expert 2010
Commented:
To answer your above (initial post) questions:
1. No...a VM with bad sectors will not affect your host disk sectors. So, if you're running VMware Server on a 2K3 box let's say, the VM's disk, which is a VMDK file, is nothing more than that...a VMDK file like a .doc file or .xls file on the host Windows OS. If there is a bad sector on the HOST disk, it may cause corruption of the files on that host disk because those files are on that disk, but not vice versa.
2. Yes, by way that "burners" shared...doing a sys recovery and/or chckdsk /r. The only other alternative depending on how bad the errors are is to just rebuild the VM.
3. If you delete the VM, the bad sectors ON THE VM will of course be gone. But if you have bad sectors on your host...no...not necessarily. It just depends on where the bad sectors are, etc. If you were to migrate a 'bad' VM to another host...beit a VMware Server host or an ESXi host, the 'bad' VM will be in tact and will work just as shaky on the new host system as it did on the original.

Regards,
~coolsport00
I avoid the /R option of chkdsk (and fix bad sectors of scandisk, exact same thing really) like the plague because both do a LOWSY job of recovering the data off the bad sectors before remapping them to spare sectors and setting them aside forever.  So yes they set aside bad sectors BUT after a handful of retries they give up much too easily and quickly trying to read the data off to be rewritten in the remapped, so, instead you end up with blocks of "zeroes" in those spots, making "holes" in the files, and essentially "corrupting" them, grrrrrrr.  You're then expected to recover from last known good backup any files/folders affected by the badsectoring process of unrecoverable sectors.  As long as you have good backups and can recover the affected files, it's sorta ok.  If the bad sectors happen on empty space no problemo.
Instead, I highly reccommend, especially for those situations where backups are out-of-date or don't exist, is GRC's SpinRite 6 http://spinrite.com  or HDDRegenerator.  SpinRite retries literally hundreds of times to read the data off the bad sectors and uses scientific signal strength variance and statistical data sampling comparison techniques to try to ascertain and recover data from the bad sectors as much as possible, often 90 or 100% data is saved.  
Just know it estimates time needed and once bad spots are encountered that ETA is overly optimistic wrong.  For speed and direct access it need the drive to be connected directly not via USB.   Interestingly SpinRite is an excellent preventative measure, identifying "iffy" sectors before they become bad thus saving your data and prolonging the useful life of your drives.  HDDRegenerator is similar.  If spinrite had irrecoverable bad blocks, you'll still want to have good backup to recover from.
The problem is once you use chkdsk /R I think it's too late for the better tool to remap the bad sectors that have already been remapped badly.
So yep, do your backups, have good backups.
Hardware RAID handle bad sectors in their own way, depending on the algorithms and techniques used by the raid and whether stright mirroring or striping or parity are used.  RAID sets aside it's own advanced bios and boot logic and extra spare sectors on the drive, and in the simplest example of mirroring if EITHER drive has a bad sector it reads the data from  the same sector of the other drive.  Some remap automatically, some require you to run the RAID bios setup utilities set once-in-a-while to scan and map badsectors in that RAID's own way.   Some enterprise class drives have ECC error correction and/or autobadsectoring.  Most consumer-level drives do not.

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial