This bug was fixed in the package libvirt - 0.9.2-4ubuntu15
---
libvirt (0.9.2-4ubuntu15) oneiric; urgency=low
* Pull patches from upstream which prevent a race between lxc driver and
controller while a container is started, easily exposed by nova.
(LP: #842845)
-- Serge
This is actually fixed upstream. I have a new independent patch (which
can be found in the thread listed in comment #17) which fixes this, but
it will be better if I can cherrypick the lxc patch from git.
** Changed in: nova (Ubuntu Oneiric)
Status: In Progress = Invalid
--
You received
** Patch added: Proposed debdiff solving the issue
https://bugs.launchpad.net/ubuntu/+source/nova/+bug/842845/+attachment/2503169/+files/libvirt-lxc-race.debdiff
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
libvirt with this debdiff attached passes qa-regression-testing
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
Title:
problems starting multiple lxc instances concurrently
To
** Patch added: debdiff solving the issue
https://bugs.launchpad.net/ubuntu/+source/nova/+bug/842845/+attachment/2504441/+files/nova-nbd.debdiff
** Changed in: libvirt (Ubuntu Oneiric)
Assignee: Serge Hallyn (serge-hallyn) = Chuck Short (zulcss)
--
You received this bug notification
@Chuck,
assigning this to you in the hopes that you can either push the debdiff
or merge it with any changes you have remaining for nova.
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
@Chuck,
sorry! wrong bug. I'm aboug to assign 861656 to you however.
** Changed in: libvirt (Ubuntu Oneiric)
Assignee: Chuck Short (zulcss) = Serge Hallyn (serge-hallyn)
** Patch removed: debdiff solving the issue
** Changed in: libvirt (Ubuntu Oneiric)
Assignee: (unassigned) = Serge Hallyn (serge-hallyn)
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
Title:
problems starting
** Tags added: rls-mgr-o-tracking
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
Title:
problems starting multiple lxc instances concurrently
To manage notifications about
** Changed in: libvirt (Ubuntu Oneiric)
Status: Invalid = In Progress
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
Title:
problems starting multiple lxc instances
Actually this libvirt debdiff may be a better solution. I'm going to
send it to upstream libvirt (cc: openstack) for comment.
** Patch added: libvirt.debdiff
https://bugs.launchpad.net/ubuntu/+source/nova/+bug/842845/+attachment/2475829/+files/libvirt.debdiff
--
You received this bug
Sent the patch upstream. The tread can be followed here:
https://www.redhat.com/archives/libvir-list/2011-September/msg01181.html
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
I'm able to reproduce this with 'euca-run-instances -n 4'. But the
libvirt instances do in fact start.
virsh -c lxc:/// list shows the, but euca-describe-instances shows them
pending.
nova-compute.log shows:
2011-09-27 16:06:23,110 ERROR nova.exception [-] Uncaught exception
(nova.exception):
I see - nova-compute checks too quickly whether the vm has powered on.
It asks libvirt to look up the vm, but libvirt complains that it has not
(yet) created the cgroups.
Not sure yet whether this should be changed at libvirt or at nova's end.
--
You received this bug notification because you
** Changed in: nova (Ubuntu Oneiric)
Assignee: (unassigned) = Chuck Short (zulcss)
** Changed in: libvirt (Ubuntu Oneiric)
Assignee: Serge Hallyn (serge-hallyn) = (unassigned)
** Changed in: nova (Ubuntu Oneiric)
Status: Incomplete = In Progress
** Changed in: libvirt (Ubuntu
With this debdiff applied, the bug no longer appears on my systems.
** Patch added: nova.debdiff
https://bugs.launchpad.net/ubuntu/+source/nova/+bug/842845/+attachment/2471726/+files/nova.debdiff
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is
The attachment nova.debdiff of this bug report has been identified as
being a patch in the form of a debdiff. The ubuntu-sponsors team has
been subscribed to the bug report so that they can review and hopefully
sponsor the debdiff. In the event that this is in fact not a patch you
can resolve
@hallyn, Can you propose your patch to upstream trunk please - then
we'll look to cherry pick it back.
Thanks :)
** Changed in: nova (Ubuntu Oneiric)
Assignee: Chuck Short (zulcss) = Serge Hallyn (serge-hallyn)
--
You received this bug notification because you are a member of Ubuntu
With the debdiff above, I can still very rarely reproduce this - it's
the same cause - nova is calling get_info while libvirt is still setting
the cgroups. The nova-compute.log shows the unhandled exception now at
_get_power_state in nova/compute/manager.py. I'm not sure why it is
getting to
Here is the relevant excerpt from nova-compute.log. note that the first
line (at /usr/lib/python2.7/dist-
packages/nova/virt/libvirt/connection.py:604)is two lines before the
timer call which should wait until the container is booted.
2011-09-27 21:20:43,830 DEBUG nova.virt.libvirt_conn [-]
In the libvirt logs, I see
21:20:43.846: 1205: debug : remoteSerializeError:132 : prog=536903814
ver=1 proc=16 type=1 serial=1004, msg=internal error Unable to get
cgroup for instance-003d
followed a bit later by
21:20:43.856: 8627: debug : virCgroupMakeGroup:511 : Make group
** Changed in: nova (Ubuntu)
Milestone: None = ubuntu-11.10
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
Title:
problems starting multiple lxc instances concurrently
To
** Changed in: libvirt (Ubuntu Oneiric)
Assignee: (unassigned) = Serge Hallyn (serge-hallyn)
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
Title:
problems starting
zul, is this still and issue and are we able to reproduce it?
Is there a nova issue?
** Also affects: libvirt (Ubuntu Oneiric)
Importance: High
Status: Confirmed
** Also affects: nova (Ubuntu Oneiric)
Importance: High
Status: New
** Changed in: libvirt (Ubuntu Oneiric)
** Tags added: server-o-rs
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
Title:
problems starting multiple lxc instances concurrently
To manage notifications about this bug
I had a look at this and this caused by libvirt in src/lxc/lxc_driver.c:
545 if (virCgroupForDomain(driver-cgroup, vm-def-name, cgroup, 0)
!= 0) {
546 lxcError(VIR_ERR_INTERNAL_ERROR,
547 _(Unable to get cgroup for %s), vm-def-name);
548
Can't seem to reproduce this with a simple
for c in o1 o2 o3 o4; do
virsh -c lxc:// start $c
done
so I'll set up an openstack box to test.
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
You don't get this problem at all doing the same thing with kvm
containers, right?
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
Title:
problems starting multiple lxc
On Mon, 2011-09-12 at 13:36 +, Serge Hallyn wrote:
You don't get this problem at all doing the same thing with kvm
containers, right?
Right
--
James Page
Ubuntu Server Developer
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to
This looks more like a problem with libvirt, as its an internal libvirt
error..
(nova.exception): TRACE: File
/usr/lib/pymodules/python2.7/nova/virt/libvirt/connection.py, line 1168, in
get_info
(nova.exception): TRACE: (state, max_mem, mem, num_cpu, cpu_time) =
virt_dom.info()
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/842845
Title:
problems starting multiple lxc instances concurrently
To manage notifications about this bug go to:
31 matches
Mail list logo
