Files related to VM get corrupted

[Jamie222]

Hello, I'm using VMWare Workstation 7.1 on WinXP. Files related to 2  virtual machines, I used recently, most likely got corrupted when I had  them running at once.

Both machines contain WinXP, with latest VMWare Tools and they're located on the same drive & partition on the host.

1. VM has corrupted these files: Disk-flat.vmdk, *.VMEM, *.vmss

2. VM has corrupted these files: Disk-flat.vmdk, *.VMEM

By corrupted I mean, that these filed cannot be copied elsewhere (on different drive).

In system log I have about 20-30 of these entries - Device \Device\Harddisk1\D has a bad block.

When I ran chkdsk on the affected drive I get about 10 errors like this:

 Deleting corrupt attribute record (128, "")
   from file record segment 92794.

(with different segment numbers)

I looked up this page on Microsoft's Knowledge base. It seems that Workstation's extensive *.lck files and directiories creation/deletion caused this.

[continuum]

Hi and welcome

when you use Workstation you produce a high load on your disks
are you sure they are still healthy ?

bad blocks usually are a hint to a disk that will probably die soon

I would make sure you have backups of any important data you have on that disks

the load produced by the lock files and cache-directories can be neglected compared with the load on the vmdks

if you do not need the cache directories - disable them - see
http://vm-sickbay.com/2010/06/01/how-to-disable-unity-for-all-vms-in-ws-71/

[WoodyZ]

if you do not need the cache directories - disable them - see

http://vm-sickbay.com/2010/06/01/how-to-disable-unity-for-all-vms-in-ws-71/

In both the aforementioned link and the read more link on that page you show the following Parameter=Value as being able to "... reduce the amount of files created in the cache -subdirectory ...".

unity.showBadges = “FALSE”
unity.showBorders = “FALSE”

Yet mdunn-vmware, a VMware employee, states in the read more link that these have no effect on the caches directory, so why do you still have them listed in that context, is the VMware employee wrong?

[Jamie222]

Yeah, the disk in question is about 2 and half months old. Smart indicates 100% health.

This happened after both machines were simultaneously running for a while - but nothing happened.

Before these machines were located on this disk, they each have been located on different drives. Should I separate them on different disks? Also the previous disks (several years old) are completely fine, they have no indication of damage.

By the way - I ran chkdsk /r on host and in phase 4 it fixed the corrupted files. Then I started up each machine (separately) and let chkdsk repair errors inside them. I've lost a few files, but I've retrieved them from backups.

I'm looking for a reliable way to prevent such problem in the future.

[miko67]

hello

I am having the same problem.

How did you run "chkdsk" on the host?

More generally, what do you use the Converter to fix a corrupted vm?

[Jamie222]

Run command-line (cmd.exe), change current drive to the one you want to check (e.g. d:), type chkdsk and press enter. Depends on which drive you want to check and if you want to fix errors, you'll get a question - most likely to disable access to given drive or plan checking on the next boot. Former is a bit risky, if you don't know what you're doing, latter is safer (unless you lose files during the check), but requires restart.

The real problem, I think, was that I had 2 VMs on the same HDD on host and had them running simultaneously. That most likely really stressed out my HDD and caused the problems. Ever since I had put one of those VMs on a different HDD, everything works smoothly.

[WoodyZ]

The real problem, I think, was that I had 2 VMs on the same HDD on host and had them running simultaneously. That most likely really stressed out my HDD and caused the problems.

I doubt that having having multiple running VM's on the same physical hard drive was the cause your issue.  I have had multiple running VM's on the same physical hard drives on dozens of system for over 7 years and have never had a problem because of this.  Also, if that were the case these forums would be inundated with posts to the same effect, which of course they'er not.

Ever since I had put one of those VMs on a different HDD, everything works smoothly.

Spreading the load across two or more physical hard disks will naturally improve performance and not just with Virtual Machines.

[Jamie222]

Look, I would really like to believe that wasn't the issue. But it was.

I bought a new HDD - 2TB. I also needed to get rid of a smaller HDD (250GB), so I shuffled data around, ended up with 2 VMs on the new HDD. Everything was working fine until I started first VM and then second without stopping the first.

The problem didn't appear immediately - I had both VMs running for a few hours, then stopped them both, and turned off the host PC. Nothing out of the ordinary. The rest of the story is in the first post.

Of course I can't rule out other reason - as you can see I had VMWare Workstation 7.1 back then. I downgraded to 6.5.4, which never gave any trouble.

[WoodyZ]

Sorry but I don't have time to debate this however it is absolutely absurd to think that running more then one virtual machine on the same physical hard drive at the same time is problematic in of and by itself.   If that were the case then there are many other scenarios where multiple read/writes are occurring with the Host OS much less the other applications and would be trashing every system out there on an ongoing and regular basis, which of course it's not happening.  Obviously better performance is gained by spreading the load across multiple hard drives and this is not just limited to Virtual Machines.

[Jamie222]

I know it sounds ridiculous. Maybe it's just my hardware and/or software setup.

Either way, I don't want to risk of losing data again, so I'll never run 2 VMs from the same HDD...