Knowledge Base

Verify the problem is specific to the virtual machine 

The problem may not be with the virtual machine but something else.   Things to clarify and verify are noted below.  Individual items noted will let you know if you should continue using this article to troubleshoot this problem.

1. Verify whether you have the same problem with more than one virtual machine on this ESX host.  If the problem is on every or most virtual machines on a host, this solution will not help you and you should look elsewhere.  It may still be helpful to finish reading this section to get some ideas of where to look.

2. Try cold migrating the virtual machine to a different ESX host and verify if the virtual machine is able to boot after cold migration.

3. Try unregistering the virtual machine and reregistering it on a different host.

4. Verify if you encounter problems accessing the storage location of the virtual machine.  Obviously this will need to be fixed before anything else with this virtual machine can be done.

5. Verify if the storage location has run out of free space.   Browsing the datastore with the virtual machine, check to see how many log files are found.  The command vdf -h will determine if the system is out of space on any storage location. 

6. Check to see if you are successfully able to browse the datastore with the virtual machine and its folder.  If there are a lot (thousands) of vmware*.log files in the folder you may need to delete the vmware*.log files using “rm vmware*.log” or if there are too many arguments for this to complete successfully you can try using “rm vmware-1[1-5]*.log” followed by “rm vmware-1*.log”, “rm vmware-2[1-5]*.log” followed by “rm vmware-2*.log”, and then “rm vmware*.log”.  This will delete all the log files for this virtual machine.

7. Check if two virtual machines are using the same .vmdk files for their disks.  In general, only one virtual machine can use a .vmdk at a time.  An exception to this are virtual machines that are clustered using MSCS.  Review the MSCS setup documentation here: http://www.vmware.com/pdf/vi3_301_201_mscs.pdf .

8. Check to see if the system is complaining about licensing.  For instance, verify if the system is licensed for ESX Starter and trying to start a virtual machine with 2 VCPUs. You may need to unlicense and license the ESX host and look further at the licensing of the host.  An unlicensed host with running virtual machines will not shut the virtual machines off for two weeks.  It will also not allow you to start a new virtual machine.  Other than that, there should be no adverse effects from unlicensing and licensing your host.

9. Verify if the virtual machine is asking for a question to be answered.   A virtual machine will stop the power on sequence at 95% sometimes after a clone through copy or the change of a SCSI adapter.  If a question is being asked, it will need to be answered on the virtual machine’s Summary page before the “power on” operation can be completed.  If the process is still not completed or no question appears to be being asked, you may need to restart the VirtualCenter Server service on your VirtualCenter Server host from within Services.  Obviously make sure VirtualCenter Server is not being used or in the middle of another active process before restarting.