yahoo-eng-team team mailing list archive
-
yahoo-eng-team team
-
Mailing list archive
-
Message #77594
[Bug 1821016] [NEW] Race on reboot, fault on spawning sriov instance after reboot
Public bug reported:
There appears to be some race between nova-compute and
neutron-sriov-agent when rebooting. When trying to bring up an
sriov-enabled instance this intermittently (but often) fails after a
fresh reboot. After restarting the neutron-sriov-agent and
nova-compute seems to fix this, i.e. can again spawn instances w/
sriov ports
* Pre-conditions:
- Sriov interfaces configured and functional, i.e. can spawn
functional sriov enabled instances
* Step-by-step reproduction steps:
- Verify sriov enabled instance can be spawned as the admin user
- Reboot compute
- Attempt to spawn an sriov enabled instance A, wait for fault
- Restart: $ sudo service neutron-sriov-agent restart ; sleep 1 ; sudo
service nova-compute restart
- Attempt to spawn an sriov enabled instance B
* Expected output:
2x ACTIVE instances A and B
* Actual output:
ERRORed instance A
nova-compute.log contains
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [req-eddf8d43-ca07-491f-931c-96b20cce7ef7 a0f1548a1b3f45379155ca1fb21c1599 7881e5796b2e4f80a9e5a7e089029bc3 - - -] [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] Failed to allocate network(s)
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] Traceback (most recent call last):
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 1939, in _build_and_run_instance
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] block_device_info=block_device_info)
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] File "/usr/lib/python2.7/dist-packages/nova/virt/libvirt/driver.py", line 2798, in spawn
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] destroy_disks_on_failure=True)
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] File "/usr/lib/python2.7/dist-packages/nova/virt/libvirt/driver.py", line 5321, in _create_domain_and_network
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] raise exception.VirtualInterfaceCreateException()
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] VirtualInterfaceCreateException: Virtual Interface creation failed
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086]
ACTIVE instance B
* Version:
Ocata running on Ubuntu xenial
neutron 10.0.7-0ubuntu1~cloud1
nova 15.1.5-0ubuntu1~cloud1
** Affects: neutron
Importance: Undecided
Status: New
** Affects: nova
Importance: Undecided
Status: New
** Tags: canonical-bootstack
** Also affects: nova
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1821016
Title:
Race on reboot, fault on spawning sriov instance after reboot
Status in neutron:
New
Status in OpenStack Compute (nova):
New
Bug description:
There appears to be some race between nova-compute and
neutron-sriov-agent when rebooting. When trying to bring up an
sriov-enabled instance this intermittently (but often) fails after a
fresh reboot. After restarting the neutron-sriov-agent and
nova-compute seems to fix this, i.e. can again spawn instances w/
sriov ports
* Pre-conditions:
- Sriov interfaces configured and functional, i.e. can spawn
functional sriov enabled instances
* Step-by-step reproduction steps:
- Verify sriov enabled instance can be spawned as the admin user
- Reboot compute
- Attempt to spawn an sriov enabled instance A, wait for fault
- Restart: $ sudo service neutron-sriov-agent restart ; sleep 1 ; sudo
service nova-compute restart
- Attempt to spawn an sriov enabled instance B
* Expected output:
2x ACTIVE instances A and B
* Actual output:
ERRORed instance A
nova-compute.log contains
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [req-eddf8d43-ca07-491f-931c-96b20cce7ef7 a0f1548a1b3f45379155ca1fb21c1599 7881e5796b2e4f80a9e5a7e089029bc3 - - -] [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] Failed to allocate network(s)
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] Traceback (most recent call last):
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] File "/usr/lib/python2.7/dist-packages/nova/compute/manager.py", line 1939, in _build_and_run_instance
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] block_device_info=block_device_info)
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] File "/usr/lib/python2.7/dist-packages/nova/virt/libvirt/driver.py", line 2798, in spawn
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] destroy_disks_on_failure=True)
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] File "/usr/lib/python2.7/dist-packages/nova/virt/libvirt/driver.py", line 5321, in _create_domain_and_network
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] raise exception.VirtualInterfaceCreateException()
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086] VirtualInterfaceCreateException: Virtual Interface creation failed
2019-03-20 13:59:36.636 22577 ERROR nova.compute.manager [instance: 48f54a52-8cb7-4963-93fa-c412954a2086]
ACTIVE instance B
* Version:
Ocata running on Ubuntu xenial
neutron 10.0.7-0ubuntu1~cloud1
nova 15.1.5-0ubuntu1~cloud1
To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1821016/+subscriptions
Follow ups