Ask Your Question
0

VM internal network broken after compute restart.

asked 2013-09-04 11:13:38 -0500

glenbot gravatar image

updated 2013-09-04 11:21:12 -0500

To preface -- I'm on Grizzly.

We had our compute node overheat and shutdown due to a cooling issue. The issue was resolved and everything came back online. The VMs were in a "Shutoff" state so I just ran "nova start uuid-of-machine" on all our machines and everything came online just fine, however now I cannot ping pre-existing vms within the internal network from a vm.

If I add a new VM then it works. I can ping it. I just can't access pre-existing VMs. I can access the VM from the floating IP though. It's like the existing VMs just dropped off the internal subnet although they have an IP associated with them. This is the first time this has happened.

What am I missing? Does something need to be flushed on the network node? If you need more information let me know.

edit retag flag offensive close merge delete

Comments

Hi glenbot..I lost the network connection when the machine was rebooted due to the loss of the gateway address I created. So each time I have lost the connection I had to set up my gateway connection again..

nishtha gravatar imagenishtha ( 2013-09-04 13:01:05 -0500 )edit

@nishtha are you going this through the interface or CLI? What commands do you run?

glenbot gravatar imageglenbot ( 2013-09-17 11:06:09 -0500 )edit

2 answers

Sort by ยป oldest newest most voted
0

answered 2013-09-04 23:30:43 -0500

You may need to manually restart the instance. go to specific compute-node where the VM seating in,

get the list of instances, and find the one you want to reboot

virsh list

destory the said instance

virsh destroy [id]

Now go to the instance's folder

cd .../instances/instance-00000XXX

There is a xml file 'libvirt.xml', # re-create the instance

virsh create libvirt.xml
virsh list
virsh dumpxml [id] > domain.xml
virsh define domain.xml
rm domain.xml

finally reboot the instance from dashboard

Hope this work for you!

edit flag offensive delete link more
0

answered 2013-09-04 11:28:50 -0500

glenbot gravatar image

Restarting all quantum services on the network box fixed the issue. Strange though that a reboot didn't.

edit flag offensive delete link more

Comments

is quantum-ovs-cleanup set to run at boot time?

darragh-oreilly gravatar imagedarragh-oreilly ( 2013-09-05 07:37:26 -0500 )edit

Im not sure if it is but I will double check this.

glenbot gravatar imageglenbot ( 2013-09-06 12:27:19 -0500 )edit

I don't think this was running at boot time. I added it to rc.local but it doesn't seem to have any effect.

glenbot gravatar imageglenbot ( 2013-09-17 10:59:47 -0500 )edit
1

Its package should have a startup script, but you might have to enable it - see https://bugzilla.redhat.com/show_bug.cgi?id=909655

darragh-oreilly gravatar imagedarragh-oreilly ( 2013-09-17 11:18:05 -0500 )edit

Thanks, everything is working now.

glenbot gravatar imageglenbot ( 2013-09-18 10:30:31 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2013-09-04 11:13:38 -0500

Seen: 1,486 times

Last updated: Sep 04 '13