yahoo-eng-team team mailing list archive
-
yahoo-eng-team team
-
Mailing list archive
-
Message #08419
[Bug 1270943] [NEW] Hypervisor crashes after instance is spawned
Public bug reported:
I am running Grizzly on Ubuntu 13.04 (so the network service ==
Quantum). Nova runs Quantum with LibvirtHybridOVSBridgeDriver and
LinuxOVSInterfaceDriver, while Quantum is configured to use GRE tunnels.
Further, Quantum runs on a dedicated node and VLAN.
Starting in mid-December, new Compute nodes that came online were unable
to spin new VMs. At the moment the nova-compute.log indicated that the
instance had spawned successfully, the hypervisor system crashed with
the following console dump message (last screen's worth):
http://pastebin.com/004MYzvR.
The installation of the Compute packages are controlled by puppet:
1:2013.1-0ubuntu2 nova-common
1:2013.1-0ubuntu2 nova-compute
1:2013.1-0ubuntu2 nova-compute-kvm
1.9.0-0ubuntu1 openvswitch-common
1.9.0-0ubuntu1 openvswitch-datapath-dkms
1.9.0-0ubuntu1 openvswitch-datapath-source
1.9.0-0ubuntu1 openvswitch-switch
1:1.0.3-0ubuntu1 python-cinderclient
1:2013.1.4-0ubuntu1 python-glance
1:0.9.0-0ubuntu1.2 python-glanceclient
1:2013.1.4-0ubuntu1.1 python-keystone
1:0.2.3-0ubuntu2.2 python-keystoneclient
1:2013.1-0ubuntu2 python-nova
1:2.13.0-0ubuntu1 python-novaclient
1:1.1.0-0ubuntu1 python-oslo.config
1:2013.1-0ubuntu2 python-quantum
1:2.2.0-0ubuntu1 python-quantumclient
1:1.3.0-0ubuntu1 python-swiftclient
1:2013.1-0ubuntu2 quantum-common
1:2013.1-0ubuntu2 quantum-plugin-openvswitch
1:2013.1-0ubuntu2 quantum-plugin-openvswitch-agent
The kernel being used is *not* controlled by Puppet and ends up being
whatever the latest and greatest version is in raring-updates. The
kernels in use: 3.8.0.34.52. I tried upgrading to 3.8.0.35.53 when it
became available, but that had no effect.
I'm lost. No idea how to debug this.
** Affects: nova
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1270943
Title:
Hypervisor crashes after instance is spawned
Status in OpenStack Compute (Nova):
New
Bug description:
I am running Grizzly on Ubuntu 13.04 (so the network service ==
Quantum). Nova runs Quantum with LibvirtHybridOVSBridgeDriver and
LinuxOVSInterfaceDriver, while Quantum is configured to use GRE
tunnels. Further, Quantum runs on a dedicated node and VLAN.
Starting in mid-December, new Compute nodes that came online were
unable to spin new VMs. At the moment the nova-compute.log indicated
that the instance had spawned successfully, the hypervisor system
crashed with the following console dump message (last screen's worth):
http://pastebin.com/004MYzvR.
The installation of the Compute packages are controlled by puppet:
1:2013.1-0ubuntu2 nova-common
1:2013.1-0ubuntu2 nova-compute
1:2013.1-0ubuntu2 nova-compute-kvm
1.9.0-0ubuntu1 openvswitch-common
1.9.0-0ubuntu1 openvswitch-datapath-dkms
1.9.0-0ubuntu1 openvswitch-datapath-source
1.9.0-0ubuntu1 openvswitch-switch
1:1.0.3-0ubuntu1 python-cinderclient
1:2013.1.4-0ubuntu1 python-glance
1:0.9.0-0ubuntu1.2 python-glanceclient
1:2013.1.4-0ubuntu1.1 python-keystone
1:0.2.3-0ubuntu2.2 python-keystoneclient
1:2013.1-0ubuntu2 python-nova
1:2.13.0-0ubuntu1 python-novaclient
1:1.1.0-0ubuntu1 python-oslo.config
1:2013.1-0ubuntu2 python-quantum
1:2.2.0-0ubuntu1 python-quantumclient
1:1.3.0-0ubuntu1 python-swiftclient
1:2013.1-0ubuntu2 quantum-common
1:2013.1-0ubuntu2 quantum-plugin-openvswitch
1:2013.1-0ubuntu2 quantum-plugin-openvswitch-agent
The kernel being used is *not* controlled by Puppet and ends up being
whatever the latest and greatest version is in raring-updates. The
kernels in use: 3.8.0.34.52. I tried upgrading to 3.8.0.35.53 when it
became available, but that had no effect.
I'm lost. No idea how to debug this.
To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1270943/+subscriptions
Follow ups
References