Friday, June 8, 2012

Migration / vmotion - of vm fails at 82%
vmotion of vm fails by throwing error -> "Source detected that destination failed to resume"

Scenario 1 (APD - "All Paths Dead issue" on either Source / Target ESX)
APD may be generally caused by improper removal of RDM's
(without removing from VM - remove/unmask at Storage end)

1) grep -i apd /var/log/vmkernel (execute on Source & Target ESX)
2) If you find any APD entries (similar to below) - your "vmkernel/COS OS"
   will busy in negotiating / trying to reheal the Dead paths and causing vMotion failures.

WARNING: NMP: nmp_DeviceAttemptFailover: Retry world failover device "naa.6090a06830772d1a80b95495e700708b"
WARNING: vmw_psp_rr: psp_rrSelectPath: Could not select path for device "naa.6090a06830772d1a80b95495e700708b"
WARNING: NMP: nmp_DeviceAttemptFailover: Retry world failover device "naa.6090a06830772d1a80b95495e700708b"

failed to issue command due to Not found (APD), try again...

Solution - # esxcfg-rescan vmhba1 && esxcfg-rescan vmhba2 (vmhbaX in your case)
Hope - This issue is resolved in ESX/ESXi 4.1 Update 1 & default with ESXi 5.0.
If no go; Unfortunately - only way to Resolve "APD issue" is restart ESX box
As the VM's does not migrate from the APD issue Host - you need downtime for all the VM's

Tip  - Take diligent mesaures while removing LUN's from Storage end (remove from OS/VM properly)

More info -

Scenario 2 (Incorrect h/w version of VM)
/var/log/vmware/hostd.log of "SOURCE ESX" contains below

ResolveCb: Failed with fault: (vmodl.fault.SystemError) {
reason = "Source detected that destination failed to resume."
msg = ""

/var/log/vmware/hostd.log of "TARGET ESX" contains below

Upgrade is required since hwVersion in config file is 3

Solution - right click on VM - upgrade virtual Hardware

Scenario 3 - UUID of NFS data store is different on source and target ESX hosts
# vdf -h (check on source and target ESX host)
if UUID is different (migration fails), generally UUID difference is caused by the way you add host to VC

you may add host by (ip / hostname / hostname.domain / FQDN)
To resolve UUID issues - follow below vmware KB..

Vmtools install failed - on windows - internal error 2318.

Sometimes You may experience a vmtools / vmware-tools installation failed,
It prompts you to uninstall existing first to continue with vmtools upgrade,
and if you try to uninstall existing vmtools - it exits with various reasons
(may happen due to existing vmtools files / registry was corrupted)

Here is the trick for you to uninstall / clean the Registry.

1) Right click on Windows VM / guest at VC - Choose install / upgrade VMware tools
2) Select Manual installation
3) Goto RDP / console of VM - ensure you find virtual CD at MyComputer
4) Find drive letter for your Virtual CD (Ex : D:) -> Go to command prompt ->
     D:\> setup.exe /c (if your OS is 32bit) or setup64.exe /c (if 64 bit OS)
5) Now you try to reinstall the VMware tools - it should proceed to install.

           Vmtools Silent installation method

Windows Guest
When it comes to mass deploy / large scale of VM's (vmware tools update) - it will be cumbersome for you to click on each VM and update VMware tools. Here is the way you can use "Silent install method" via SSH log on script / Power CLI.

To perform a silent, non GUI with suppressed reboot VMware Tools installation in a Windows guest operating system:

Run the command:

setup.exe /S /v /qn REBOOT=R

Note: The installer might indicate if a reboot is necessary by exiting with ERROR_SUCCESS_REBOOT_REQUIRED.

Alternatively, in vCenter Server, right-click on a virtual machine, click Install/Upgrade VMware Tools, and enter /S /v /qn REBOOT=R the Advanced field.

Source - VMware KB --> 1018377