Issue:
You receive the following errors when attempting to power on a VM.
A file was not found [lun_name] vm_name-######.vmdk
Cannot open the disk ‘/vmfs/volumes/{GUID}/vm_name/vm_name.vmdk’
Snapshot hierarchy is too deep.
Cause:
Too many snapshots have been applied to the VM. This is usually caused by automated processes that create snapshots but fail to properly delete them, but can be caused by manual snapshot creation as well.
Solution:
Method 1 – Copy disks, do not commit snapshots
This method copies the original disk (VMDK files) to a new VM, without committing the snapshots. This will allow you to work on recovering the original VM as needed.
1) Record the following VM’s settings from the original VM.
a. Amount of RAM
b. Number of CPUs
c. Number of NICs
d. Network Adapter Type (flexible or enhanced)
e. Network Label
f. Disk sizes
i. Needed to determine space needs for new server.
g. Disk names and locations
2) Verify which LUNs have the needed free space for the new VM.
3) Create a new VM using the VI Client
a. Apply the same settings are recorded earlier
b. When asked, do not create a disk
c. Record which LUN you created the new VM on.
4) Use the vmkfstools command to copy the original disks to the new location
a. On the ESX server console type the following
cd /vmdk/vmfs/volumes/lun_name/vm_name
vmkfstools -i vm_name.vmdk /vmfs/volumes/lun_name/vm_name-new/vm_name-new.vmdk
b. Repeat for each disk/vmdk file
5) Attach disks to new VM
a. Edit properties of new VM
b. Click Add
c. Select Hard Disk
i. Click Next
d. Select use an existing virtual disk
i. Click Next
e. Browse to new vmdk file
i. Click Open
ii. Click Next
f. Configure Virtual device node (SCSI ID) to match original server
i. Click Next
g. Click Finish
h. Click OK
i. Repeat for each disk.
6) Power on new VM
Method 2 – Commit snapshots
This method will commit the snapshots, 30 at a time, to the original disks. This can be very time consuming and does require 2-3 times the original size of disks in free space. For example, if you had a VM with 2 disks totally 50GB, with 30+ snapshots, you would need between 100GB and 150GB of free space.
Before stating this procedure backup the files a needed.
Committing snapshots fails with the error: Too many levels of redo logs
Method 3 – Delete all snapshots
This method will delete all snapshots and recover only the original disk. This method is similar to method 1, but you will loose any chance of recovering the snapshots later.
1) Edit the VM Properties
2) Remove all hard drives
a. Do not delete the hard drives from disk.
3) Add the hard drives you just removed by browsing to the base disk.
4) Delete snapshot vmdk files as needed using the following command.
5) Power on the VM.