[lxc-users] LXD uptime back to 0

Wed Mar 23 02:10:13 UTC 2016

D'oh, I was misremember how it works.  And good thing too, as what I was
thinking couldn't possibly work.  The number which should be used in
/proc/uptime is the st.st_ctime for /proc/<pid>.

Quoting Benoit GEORGELIN - Association Web4all (benoit.georgelin at web4all.fr):
> For container "ben" 
> 
> This would be : 
> 
> Dir: /sys/fs/cgroup/systemd/lxc/ben 
> File: tasks 
> 
> /sys/fs/cgroup/systemd/lxc/ben] stat tasks 
> File: ‘tasks’ 
> Size: 0 Blocks: 0 IO Block: 4096 regular empty file 
> Device: 17h/23d Inode: 439 Links: 1 
> Access: (0664/-rw-rw-r--) Uid: ( 0/ root) Gid: (165536/ UNKNOWN) 
> Access: 2016-03-15 01:45:44.239558654 +0100 
> Modify: 2016-03-22 01:45:43.368945755 +0100 
> Change: 2016-03-22 01:45:43.368945755 +0100 
> Birth: - 
> 
> Dir: /sys/fs/cgroup/systemd/lxc/ben/init.scope 
> File: tasks 
> 
> 
> /sys/fs/cgroup/systemd/lxc/ben/init.scope] stat tasks 
> File: ‘tasks’ 
> Size: 0 Blocks: 0 IO Block: 4096 regular empty file 
> Device: 17h/23d Inode: 445 Links: 1 
> Access: (0644/-rw-r--r--) Uid: (165536/ UNKNOWN) Gid: (165536/ UNKNOWN) 
> Access: 2016-03-15 01:45:44.415562979 +0100 
> Modify: 2016-03-15 01:45:44.415562979 +0100 
> Change: 2016-03-15 01:45:44.415562979 +0100 
> Birth: - 
> 
> 
> Or after lxc info ben --verbose 
> 
> lxc info ben --verbose 
> Name: ben 
> Architecture: x86_64 
> Created: 2016/03/07 23:39 UTC 
> Status: Running 
> Type: persistent 
> Profiles: default 
> Pid: 19782 
> Processes: 21 
> 
> Then : 
> 
> /sys/fs/cgroup/systemd/lxc/ben/init.scope] stat /sys/fs/cgroup/systemd/`awk -F: '/systemd/ { print $3 }' /proc/19782/cgroup` 
> File: ‘/sys/fs/cgroup/systemd//lxc/ben/init.scope’ 
> Size: 0 Blocks: 0 IO Block: 4096 directory 
> Device: 17h/23d Inode: 441 Links: 2 
> Access: (0755/drwxr-xr-x) Uid: (165536/ UNKNOWN) Gid: (165536/ UNKNOWN) 
> Access: 2016-03-15 01:45:44.415562979 +0100 
> Modify: 2016-03-15 01:45:44.415562979 +0100 
> Change: 2016-03-15 01:45:44.415562979 +0100 
> Birth: - 
> 
> 
> 
> Thanks 
> 
> Cordialement, 
> 
> Benoît 
> Afin de contribuer au respect de l'environnement, merci de n'imprimer ce mail qu'en cas de nécessité 
> 
> 
> De: "Serge Hallyn" <serge.hallyn at ubuntu.com> 
> À: "lxc-users" <lxc-users at lists.linuxcontainers.org> 
> Envoyé: Mardi 22 Mars 2016 19:36:09 
> Objet: Re: [lxc-users] LXD uptime back to 0 
> 
> Oh, right, but what we really want is the cgroups 
> of the container init task (lxc-info -n vps-01 -p -H) 
> 
> Quoting Benoit GEORGELIN - Association Web4all (benoit.georgelin at web4all.fr): 
> > I cannot hide anything then ! 
> > You are right . I use openvswitch 
> > 
> > 
> > cat /proc/1942/cgroup 
> > 10:blkio:/system.slice/lxd.service 
> > 9:memory:/system.slice/lxd.service 
> > 8:hugetlb:/system.slice/lxd.service 
> > 7:perf_event:/system.slice/lxd.service 
> > 6:cpuset:/system.slice/lxd.service 
> > 5:devices:/system.slice/lxd.service 
> > 4:cpu,cpuacct:/system.slice/lxd.service 
> > 3:freezer:/system.slice/lxd.service 
> > 2:net_cls,net_prio:/system.slice/lxd.service 
> > 1:name=systemd:/system.slice/lxd.service 
> > 
> > 
> > 
> > stat /sys/fs/cgroup/systemd/`awk -F: '/systemd/ { print $3 }' /proc/1942/cgroup` 
> > File: ‘/sys/fs/cgroup/systemd//system.slice/lxd.service’ 
> > Size: 0 Blocks: 0 IO Block: 4096 directory 
> > Device: 17h/23d Inode: 144 Links: 2 
> > Access: (0755/drwxr-xr-x) Uid: ( 0/ root) Gid: ( 0/ root) 
> > Access: 2016-03-23 00:20:41.844854933 +0100 
> > Modify: 2016-03-23 00:20:41.844854933 +0100 
> > Change: 2016-03-23 00:20:41.844854933 +0100 
> > Birth: - 
> > 
> > Cordialement, 
> > 
> > Benoît 
> > Afin de contribuer au respect de l'environnement, merci de n'imprimer ce mail qu'en cas de nécessité 
> > 
> > 
> > De: "Serge Hallyn" <serge.hallyn at ubuntu.com> 
> > À: "lxc-users" <lxc-users at lists.linuxcontainers.org> 
> > Envoyé: Mardi 22 Mars 2016 19:17:39 
> > Objet: Re: [lxc-users] LXD uptime back to 0 
> > 
> > Quoting Benoit GEORGELIN - Association Web4all (benoit.georgelin at web4all.fr): 
> > > Interesting, look the process monitor : 
> > > 
> > > root 1942 0.0 0.1 72544 1280 ? Ss Mar10 0:00 [lxc monitor] /var/lib/lxd/containers vps-01 
> > > root 1984 0.0 0.0 72544 948 ? S Mar10 0:00 \_ [lxc monitor] /var/lib/lxd/containers vps-01 
> > > root 19734 0.0 0.3 72544 3460 ? Ss Mar15 0:00 [lxc monitor] /var/lib/lxd/containers vps-02 
> > > root 19781 0.0 0.2 72544 2364 ? S Mar15 0:00 \_ [lxc monitor] /var/lib/lxd/containers vps-02 
> > 
> > Hi, 
> > 
> > I bet you're using openvswitch? That causes lxc to create a second 
> > thread which waits for the container to stop and deletes the port. 
> > 
> > > They exist twice for each containers 
> > > 
> > > init process : 
> > > 
> > > 165536 1987 2.3 0.2 28280 3036 ? Ss Mar10 429:30 \_ /sbin/init 
> > > 165536 19782 0.0 0.4 37388 4476 ? Ss Mar15 0:04 \_ /sbin/init 
> > > 
> > > Where can I get the ctime of the cgroups of the tatks and the init task ? 
> > 
> > cat /proc/1942/cgroup 
> > stat /sys/fs/cgroup/systemd/`awk -F: '/systemd/ { print $3 }' /proc/1942/cgroup` 
> > _______________________________________________ 
> > lxc-users mailing list 
> > lxc-users at lists.linuxcontainers.org 
> > http://lists.linuxcontainers.org/listinfo/lxc-users 
> 
> > _______________________________________________ 
> > lxc-users mailing list 
> > lxc-users at lists.linuxcontainers.org 
> > http://lists.linuxcontainers.org/listinfo/lxc-users 
> 
> _______________________________________________ 
> lxc-users mailing list 
> lxc-users at lists.linuxcontainers.org 
> http://lists.linuxcontainers.org/listinfo/lxc-users 

> _______________________________________________
> lxc-users mailing list
> lxc-users at lists.linuxcontainers.org
> http://lists.linuxcontainers.org/listinfo/lxc-users