VFS: Cannot open root device "LABEL=cloudimg-rootfs" or unknown-block(0,0): error -6
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Expired
|
Undecided
|
Unassigned | ||
linux (Ubuntu) |
Expired
|
Undecided
|
Unassigned |
Bug Description
Description
===========
An kvm instance failed to boot with a kernel panic after hitting this error:
VFS: Cannot open root device "LABEL=
Steps to reproduce
==================
This doesn't reproduce reliably, it seems to be a race condition.
* I started an instance through the API. It had not attached storage, and was attached to a single network. It used an Ubuntu Xenial cloud image which booted successfully on a different instance on a different compute node.
Expected Result
===============
* I expected the instance to boot successfully
Actual Result
=============
* The instance failed to boot with a kernel panic:
http://
* I used virsh destroy to stop that instance, then started it again through the API, and it worked.
Environment
===========
xenial/
ii nova-common 2:14.0.
ii nova-compute 2:14.0.
ii nova-compute-kvm 2:14.0.
ii nova-compute-
ii python-nova 2:14.0.
ii python-novaclient 2:6.0.0-
ii libvirt-bin 1.3.1-1ubuntu10.6 amd64 programs for the libvirt library
ii libvirt0:amd64 1.3.1-1ubuntu10.6 amd64 library for interfacing with different virtualization systems
ii nova-compute-
ii python-libvirt 1.3.1-1ubuntu1 amd64 libvirt Python bindings
Changed in nova: | |
status: | New → Incomplete |
Seems like a race in populating and using the labels on the disks (or an I/O error, but I haven't found one in the log).
In any case I think this should get to the attention of the kernel Team.