On Wed, Feb 01, 2017 at 07:13:25PM +0900, Tomasz Chmielewski wrote:
> FYI it was crashing in the following conditions:
>
> - privileged container (i.e. the one for docker)
> - on btrfs filesystem
> - btrfs subvolume created inside the container (docker would create such
> subvolumes)
> - snapshot
FYI it was crashing in the following conditions:
- privileged container (i.e. the one for docker)
- on btrfs filesystem
- btrfs subvolume created inside the container (docker would create such
subvolumes)
- snapshot of the container made
Will your fix eventually go to i.e. 2.0.9?
Tomasz
O
Hey there,
I wrote a fix for that function just yesterday which I think should fix
your issue. It's been merged in git but isn't in any released version of
LXD yet.
Since you're using LXD 2.0.8, I made a build of 2.0.8 with that one fix
applied at: https://dl.stgraber.org/lxd-2.0.8-btrfs
SHA256:
Unfortunately it's still crashing, around 1 day after removing the
docker container:
Feb 1 06:16:20 lxd01 lxd[2593]: error: LXD still not running after 600s
timeout.
Feb 1 06:16:20 lxd01 systemd[1]: lxd.service: Control process exited,
code=exited status=1
Feb 1 06:16:20 lxd01 systemd[1]:
I think it may be related to
https://www.stgraber.org/2016/04/13/lxd-2-0-docker-in-lxd-712/
I have a docker container, with several dockers inside, and with lxd
snapshots.
Doing this:
# lxc delete docker
error: Get
http://unix.socket/1.0/operations/7d30bf41-3af6-4b48-b42c-06fdd2bba48b/wait:
lxd process on one of my servers started to hang a few days ago.
In syslog, I can see the following being repeated (log below).
I see a similar issue here:
https://github.com/lxc/lxd/issues/2089
but it's closed.
Running:
ii lxd 2.0.8-0ubuntu1~ubuntu16.04.2