Virtual Machine Reboots and Filesystem Corruption as the Virtual Machine is Running on more than one Hypervisor, Error: "<VM: xxxx, had been running on server: xxxx, but is now running on server: xxxx>" (Doc ID 2498430.1)

Last updated on JUNE 04, 2020

Applies to:

Symptoms

Unexpected virtual machine reboot and filesystem corruption with running virtual machine.

The logs on one Oracle VM Server show that the VM's were live migrated to other Oracle VM Servers.

From the ovs-agent.log:

[2019-01-15 16:28:49 8119] DEBUG (ha:97) VM 0004fb00000600001234567890abcdef started on <OVS1_IP_address>
[2019-01-15 16:28:49 8119] DEBUG (ha:97) VM 0004fb00000600002234567890abcdef started on <OVS2_IP_address>
[2019-01-15 16:28:49 8119] DEBUG (ha:97) VM 0004fb00000600003234567890abcdef started on <OVS2_IP_address>
[2019-01-15 16:28:49 8119] DEBUG (ha:97) VM 0004fb00000600004234567890abcdef started on <OVS4_IP_address>

But  "xm list" on all Oracle VM Servers show the same VM is running on more that one server.

In the AdminServer.log:

####<2019-01-15T16:28:02.518+0300> ... <VM: vm40, had been running on server: vmhost03, but is now running on server: vmhost01>  
####<2019-01-15T16:28:02.529+0300> ... <VM: vm06, had been running on server: vmhost03, but is now running on server: vmhost02>  
####<2019-01-15T16:28:03.546+0300> ... <VM: vm46, had been running on server: vmhost03, but is now running on server: vmhost02>  
####<2019-01-15T16:28:12.685+0300> ... <VM: vm32, had been running on server: vmhost03, but is now running on server: vmhost04> 

Changes

 To solve a previous issue the DLM lock in /etc/xen/xend-config.sxp were commented out on vmhost03

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.