failed restoreThere has been a known bug in VMware ESX’s Changed Block Tracking (CBT) for some time now. Veeam released a workaround for this problem back in November but VMware hasn’t patched the issue until recently.

Basically, if you expanded a virtual disk with CBT past 128 GB, the data was not backed up completely and got corrupted. If you ever had a failed restore and couldn’t figure out why, this CBT bug might have had something to do with it.

VMware has released patches for all versions of ESX over 5. Here are the patches:

Currently, there are no patches for this problem in ESXi 4 because it is no longer supported. To avoid corrupted backup data, you’ll have to disable CBT and the enable it for disks extended past the 128 GB boundary. Do this through a vCenter Client only on a powered off virtual machine.

To disable CBT and then enable it for disks extended past a128 GB boundary using the vCenter Client:

  1. Power off the virtual machine.
  2. Use a vCenter client to disable CBT on each disk extended past a 128 GB boundary.
  3. Power on the virtual machine.
  4. Power off the virtual machine.
  5. Use a vCenter client to enable CBT on each disk where you disabled it in step 2.
  6. Power on the virtual machine.

To work around this issue where the virtual machine must remain powered on, perform the following steps:

  1. Disable CBT using either PowerCLI or the VDDK API.
  2. Take a snapshot.
  3. Delete the snapshot (To recover space and performance).
  4. Enable CBT using either PowerCLI or the VDDK API.
  5. Take a snapshot.
  6. Delete the snapshot.

The next backup after toggling CBT is the full backup of the virtual machine.

Note: Discard any backups that were captured after growing the disk because they might be incomplete.

Visit VMware’s Knowledge Base article for more detailed information.

Have questions about VMware or virtual backups? Email us or give us a call: 502-240-0404!