Better error when no virtualization method is available on the server
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Compute (nova) |
Fix Released
|
Low
|
Josh Kearney |
Bug Description
When no hardware virtualisation is available (which is used on the default install) on the server and trying to run an instance nothing are get back to the user and the logs contains only an error like this :
==> /var/log/
2010-10-13 16:41:13+0100 [-] (root): ERROR instance 4014792697: Failed to spawn
2010-10-13 16:41:13+0100 [-] Traceback (most recent call last):
2010-10-13 16:41:13+0100 [-] File "/usr/lib/
2010-10-13 16:41:13+0100 [-] yield self.driver.
2010-10-13 16:41:13+0100 [-] File "/usr/lib/
2010-10-13 16:41:13+0100 [-] result = g.send(result)
2010-10-13 16:41:13+0100 [-] File "/usr/lib/
2010-10-13 16:41:13+0100 [-] yield self._conn.
2010-10-13 16:41:13+0100 [-] File "/usr/lib/
2010-10-13 16:41:13+0100 [-] if ret is None:raise libvirtError(
2010-10-13 16:41:13+0100 [-] libvirtError: internal error no supported architecture for os type 'hvm'
2010-10-13 16:41:13+0100 [-] Unhandled error in Deferred:
2010-10-13 16:41:13+0100 [-] Unhandled Error
Traceback (most recent call last):
A more meaningful error would be nice for the users to enable virtualisation in BIOS.
Related branches
- Jay Pipes (community): Approve
- Matt Dietz (community): Approve
-
Diff: 52 lines (+12/-5)2 files modifiednova/compute/manager.py (+11/-3)
plugins/xenserver/xenapi/etc/xapi.d/plugins/glance (+1/-2)
Changed in nova: | |
status: | Confirmed → In Progress |
assignee: | nobody → Josh Kearney (jk0) |
Changed in nova: | |
status: | In Progress → Fix Committed |
Changed in nova: | |
milestone: | none → 2011.2 |
status: | Fix Committed → Fix Released |
is it libvirt's "no supported architecture for os type 'hvm'" that you think should be better, or are you saying we should catch that error and translate it to something more readable? The latter seems like a rabbit hole I don't want to go down.