[lxc-users] cgmanager_move_pid_abs_sync failed: invalid request

[Benoit GEORGELIN - Association Web4all]

Hi everyone, 

Can someone explain to me why/how this problem can happen : LXC is running for days, suddenly something happen and LXC is not able to communicate with some process related to cgmanager . Only option is to stop/start the container again 

lxc-ls -f 

NAME STATE IPV4 IPV6 GROUPS AUTOSTART 
------------------------------------------------------------- 
yunohost RUNNING xxx.xxx.xxx.xxx - - NO 

lxc-attach -n yunohost 

lxc-attach: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-attach: cgmanager.c: cgm_attach: 1460 Failed to enter group /user.slice 
lxc-attach: attach.c: lxc_attach: 909 error communicating with child process 

lxc-info -n yunohost 

lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 
Name: yunohost 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 
State: RUNNING 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 
PID: 26614 
IP: xxx.xxx.xxx.xxx 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup cpuacct: 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup blkio: 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup memory: 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 
lxc-info: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-info: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup memory: 
Link: vethJCTHVT 
TX bytes: 10.89 MiB 
RX bytes: 390.76 MiB 
Total bytes: 401.65 MiB 


benoit at lxd-virt-01a:~$ lxc-stop -n yunohost 
lxc-stop: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-stop: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 
lxc-stop: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-stop: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 
lxc-stop: cgmanager.c: lxc_cgmanager_enter: 698 call to cgmanager_move_pid_abs_sync failed: invalid request 
lxc-stop: cgmanager.c: do_cgm_get: 901 Failed to enter container cgroup freezer: 

lxc-ls -f 
NAME STATE IPV4 IPV6 GROUPS AUTOSTART 
------------------------------------------------------------- 
yunohost STOPPED - - - NO 


lxc-start -n yunohost 

lxc-attach -n yunohost 
root at yunohost:/# uptime 
19:27:35 up 0 min, 0 users, load average: 0,74, 0,69, 0,68 

I'm missing some parts to understand what means this error. I'm running an unprivileged container 
It happened to me at least 2 to 3 times since I'm using LXC/LXD 

I'm not able yet to explain how to reproduce. 
Does "service lxc restart" or "service lxd restart" or "service cgmanager restart" can introduce this ? 

I feel like it's only related to systemd and dbus since I cannot have a "normal" debian 8 system working well with dbus, avhi-daemon and stuff like that. 


All LXC containers are running 
Is there any thing I can do but stop and start the container ? 
What is technically missing from lxc to work ? 

Thanks a lot for your help . 

Have a nice (long?) weekend 

Cordialement, 

Benoît Georgelin 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxcontainers.org/pipermail/lxc-users/attachments/20150905/28d9d6a6/attachment.html>