[lxc-users] lxcfs segfaults since around 2019-07-23, containers malfunction

[Stéphane Graber]

This has now been resolved in the stable snap.

Affected users should:
 - snap refresh lxd
 - ps aux | grep lxcfs (to confirm lxcfs is now running)
 - lxc restart <affected containers>
 - If a container restart isn't an option, then "grep lxcfs
/proc/mounts" inside the container and proceed to "umount" all of
those

More details on this issue can be found here:
https://github.com/lxc/lxcfs/issues/295

On Tue, Jul 23, 2019 at 9:44 AM Christian Brauner <christian at brauner.io> wrote:
>
> On Tue, Jul 23, 2019 at 01:31:12PM +0200, Tomasz Chmielewski wrote:
> > Since around 2019-07-23, lxcfs segfaults randomly on Ubuntu 18.04 servers
> > with LXD from snap:
> >
> > lxcfs[1424]: segfault at 0 ip 00007f518f5e4326 sp 00007f519da1f9a0 error 4
> > in liblxcfs.so[7f518f5d8000+1a000]
> >
> >
> > As a result, containers malfunction. If a container is stopped, then started
> > again - everything works well, however - after a few hours, it happens
> > again.
> >
> > root at uni01:~# free
> > Error: /proc must be mounted
> >   To mount /proc at boot you need an /etc/fstab line like:
> >       proc   /proc   proc    defaults
> >   In the meantime, run "mount proc /proc -t proc"
> >
> > root at uni01:~# uptime
> > Error: /proc must be mounted
> >   To mount /proc at boot you need an /etc/fstab line like:
> >       proc   /proc   proc    defaults
> >   In the meantime, run "mount proc /proc -t proc"
> >
> >
> > root at uni01:~# ls /proc
> > ls: cannot access '/proc/stat': Transport endpoint is not connected
> > ls: cannot access '/proc/swaps': Transport endpoint is not connected
> > ls: cannot access '/proc/uptime': Transport endpoint is not connected
> > ls: cannot access '/proc/cpuinfo': Transport endpoint is not connected
> > ls: cannot access '/proc/meminfo': Transport endpoint is not connected
> > ls: cannot access '/proc/diskstats': Transport endpoint is not connected
> > (...)
> >
> > Is it a known issue? I'm observing it on around 10 servers.
>
> I've had that issue reported today. Can you get a coredump for this so I
> can see where lxcfs is segfaulting?
>
> Christian
> _______________________________________________
> lxc-users mailing list
> lxc-users at lists.linuxcontainers.org
> http://lists.linuxcontainers.org/listinfo/lxc-users



-- 
Stéphane