What are the causes let OpenNebula "thinks" VM is "POWEROFF" and how to recover?

Community Support
1

Hi Everyone,

Anyone had ever encounter an issue about VM stat changes to “POWEROFF” a few minutes (2 ~ 3 minutes) after “RUNNING”?

The actual problem is that the “POWEROFF” stat is incorrect as the VM is actually running on the computing node. If you try to “boot” it back, it will failed as that instance is already running on the same computing node.

Is it a bug of OpenNebula? Any suggestions to get it recovered?

Thanks,

Gene

************************* VM log****************************************

Thu Feb 12 14:46:36 2015 [Z0][DiM][I]: New VM state is ACTIVE.
Thu Feb 12 14:46:37 2015 [Z0][LCM][I]: New VM state is PROLOG.
Thu Feb 12 14:49:43 2015 [Z0][LCM][I]: New VM state is BOOT
Thu Feb 12 14:49:43 2015 [Z0][VMM][I]: Generating deployment file:
/var/lib/one/vms/42922/deployment.0
Thu Feb 12 14:49:43 2015 [Z0][VMM][I]: ExitCode: 0
Thu Feb 12 14:49:43 2015 [Z0][VMM][I]: Successfully execute
network driver operation: pre.
Thu Feb 12 14:49:46 2015 [Z0][VMM][I]: ExitCode: 0
Thu Feb 12 14:49:46 2015 [Z0][VMM][I]: Successfully execute
virtualization driver operation: deploy.
Thu Feb 12 14:49:46 2015 [Z0][VMM][I]: ExitCode: 0
Thu Feb 12 14:49:46 2015 [Z0][VMM][I]: Successfully execute
network driver operation: post.
Thu Feb 12 14:49:46 2015 [Z0][LCM][I]: New VM state is RUNNING
Thu Feb 12 14:51:48 2015 [Z0][DiM][I]: New VM state is POWEROFF

2

This should be solved in one 4.10.2 what version are you using?

3

Hi Ruben,

Thank you for your reply!
We are on 4.8.0.

Gene

4

BTW, is there a way to patch the fix? We do not plan upgrade to 4.10.2 shortly.

Thanks, Gene

5

Sorry, this is not an easy one :frowning:

You are interested in the commits referenced by this issue:

http://dev.opennebula.org/issues/3212

1 Like
6

Thank Ruben for your clarification!

We will consider pull up the upgrade because this bug is really something blocking our “automation”.

Gene