** Changed in: lxd (Ubuntu)
Status: Incomplete => Invalid
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1667444
Title:
Failed to start lxd container because it is already running a start
** Changed in: lxd (Ubuntu)
Status: New => Incomplete
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1667444
Title:
Failed to start lxd container because it is already running a start
@jjohansen,
Thanks for looking into this and providing the debugging suggestions. It
will take a little bit of time to setup things to inject this into our
automated testing. But hopefully I'll have this enabled soon and we
we'll keep monitoring for a repeat.
--
You received this bug
Looking through the logs in comment 8 for any indications of restarts,
unloads, I came up with this interesting discrepancy. Basically the
failure is recorded as happening 37s before the profile is loaded.
The load message for the container's profile
Feb 18 14:06:11 elkhart kernel: [ 494.893832]
Note: I did find messages being lost so disabling printk rate limiting
is really important
eg
Feb 18 14:06:56 elkhart kernel: [ 539.725207] audit_printk_skb: 15 callbacks
suppressed
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
it would also be good to get the output of the apparmor profiles file,
before and after a failed run
cat /sys/kernel/security/apparmor/profiles
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1667444
>From what I can see in the logs this does not appear to be the issues
Tyler and Stephane have mentioned, however I can't entirely rule that
issue out yet.
The logs in comment 1 do not match up with the error reported in the bug
description, but appear to be the same as from comment 8, however
This may be a known issue that we're currently discussing privately.
I'll update this bug when there's more public info available.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1667444
Title:
Pinged jjohansen to take a look whether he has seen this issue before.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1667444
Title:
Failed to start lxd container because it is already running a
So this could be a potential kernel bug I'm not sure what LXD would do
that could cause this error. We should ping someone from the AppArmor
team and see if they have an ideas as to how this could be caused.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is
And here are the logs from ci build 5481. The failure was with lxc
5-lxd-1.
** Attachment added: "lp-1667444-5481.tar.gz"
https://bugs.launchpad.net/ubuntu/+source/lxd/+bug/1667444/+attachment/4838315/+files/lp-1667444-5481.tar.gz
--
You received this bug notification because you are a
Here is another set of logs from ci run 5477. The lxc 2-lxd-6 is the one
that failed to start.
** Attachment added: "lp-1667444-5477.tar.gz"
https://bugs.launchpad.net/ubuntu/+source/lxd/+bug/1667444/+attachment/4838314/+files/lp-1667444-5477.tar.gz
--
You received this bug notification
Here is all of /var/log for the host machine from build 5330.
** Attachment added: "lp-1667444-5330.tar.gz"
https://bugs.launchpad.net/ubuntu/+source/lxd/+bug/1667444/+attachment/4838296/+files/lp-1667444-5330.tar.gz
--
You received this bug notification because you are a member of Ubuntu
I mean /var/log/lxd//lxc.log
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1667444
Title:
Failed to start lxd container because it is already running a start
operation
To manage notifications
@fginther, it would be cool to see the dmesg output, /var/log/lxd,
/var/log/lxd//forkstart.log, /var/log/lxd/lxc.log. I
suspect, like Stéphane that this might be an AppArmor bug we've been
seeing quite regularly.
--
You received this bug notification because you are a member of Ubuntu
Bugs,
@stgraber
> What's that environment like? Are all those containers running on the host
> directly or are they nested?
These containers are all running directly on the host (an amd64 box).
The host is managed by maas with juju driving the deployment and juju
creating the lxcs. The host is running
I did more tests. In fact, it is working fine when the node where you
want to create a container is in 16.04 (Xenial). When the node is 14.04
(trusty), it is failing. I can see it is trying to build a Xenial
container onto a trusty OS node.
Hereunder an extract of juju status command where you
Hello. I have the same problem trying to deploy a charm with juju on a
node through an LXD container. All (MAAS, Juju, Nodes) running onto IBM
Power platform (ppc64le).
Hereunder an extract of the lxd.log of my juju node:
ephemeral=false lvl=info msg="Creating container" name=juju-fc091e-44-lxd-3
So the error suggests that the needed apparmor profile couldn't be found
at the time the container started but there's nothing very definitive in
any of the logs as to why that would be.
It reminds me of an occasional kernel bug that we've been running into
with regard to apparmor namespacing
19 matches
Mail list logo
