[lxc-devel] Priorities for LXC 1.0.1

[S.Çağlar Onur]

On Thu, Mar 6, 2014 at 10:19 AM, Serge Hallyn <serge.hallyn at ubuntu.com> wrote:
> Quoting Serge Hallyn (serge.hallyn at ubuntu.com):
>> Quoting S.Çağlar Onur (caglar at 10ur.org):
>> > On Wed, Mar 5, 2014 at 11:16 PM, S.Çağlar Onur <caglar at 10ur.org> wrote:
>> > > On Wed, Mar 5, 2014 at 11:08 PM, Serge Hallyn <serge.hallyn at ubuntu.com> wrote:
>> > >> Quoting S.Çağlar Onur (caglar at 10ur.org):
>> > >>> On Wed, Mar 5, 2014 at 3:58 PM, Stéphane Graber <stgraber at ubuntu.com> wrote:
>> > >>> > On Wed, Mar 05, 2014 at 10:38:49AM -0500, Stéphane Graber wrote:
>> > >>> >> On Tue, Feb 25, 2014 at 07:53:06PM -0500, Stéphane Graber wrote:
>> > >>> >> > Hello,
>> > >>> >> >
>> > >>> >> > Ideally, I'd like to have LXC 1.0.1 out in the next week or so with all
>> > >>> >> > of the fixes we stacked since the 1.0 release last month.
>> > >>> >> >
>> > >>> >> > On top of those, the following issues have been mentioned a few times
>> > >>> >> > and we probably need to fix those before 1.0.1 too:
>> > >>> >> >  - unprivileged containers spawned by root are broken.
>> > >>> >> >  - lxc-attach on systems with rshared / is broken.
>> > >>> >> >  - the mount table of containers spawned on systems with rshared / is
>> > >>> >> >    poluted (twice as long as it should be).
>> > >>> >>
>> > >>> >>
>> > >>> >> All of those issues and more have now been resolved with fixes available
>> > >>> >> in the stable-1.0 branch.
>> > >>> >>
>> > >>> >> Unless someone comes up with a major issue by then, I intend to release
>> > >>> >> 1.0.1 later this afternoon as our first LXC 1.0 bugfix release
>> > >>> >> (admitedly a rather massive bugfix release for our first one...).
>> > >>> >
>> > >>> >
>> > >>> > Caglar reported a problem with lxc-test-concurent and similar use from
>> > >>> > Go related to threading and cgmanager interactions, I'm postponing 1.0.1
>> > >>> > until we have this resolved.
>> > >>>
>> > >>> Still testing but looks like following patch helped (on top of current master)
>> > >>
>> > >> Thanks.  Since this will be called during cgroup_ops_init()
>> > >> which has the attribute constructor, this will be done in
>> > >> time.
>> > >>
>> > >> Needs your sign-off, but
>> > >
>> > > Just sent to the list but let me play with it for a while before
>> > > applying as I only tested the Go side till now.
>> >
>> > Unfortunately, looks like this patch is not playing well with
>> > lxc-test-concurrent (but seems to be working fine with it's Go
>> > counterpart). It got stuck as soon as it tries to start containers :/
>> >
>> > sudo ./lxc-test-concurrent
>> >
>> > Iteration 1/1 maxfd:6
>> > Executing (create) for 5 containers...
>> > Executing (start) for 5 containers...
>> >
>> > It could be because of dbus 1.6 as [1] says following;
>> >
>> > """
>> > If this function is not called, the D-Bus library will not lock any
>> > data structures. If it is called, D-Bus will do locking, at some cost
>> > in efficiency.
>> >
>> > Since D-Bus 1.7 it is safe to call this function from any thread, any
>> > number of times (but it must be called before any other libdbus API is
>> > used).
>> >
>> > In D-Bus 1.6 or older, this function must be called in the main thread
>>
>> D'oh.  I missed that we're running 1.6.  Ok, no problem,
>> just add a 'static bool dbus_threads_initialized = false' at top, then do
>>
>>       if (!dbus_threads_initialized) {
>>               dbus_threads_init_default();
>>               dbus_threads_initialized = true;
>>       }
>>
>> :)
>>
>> > before any other thread starts. As a result, it is not sufficient to
>> > call this function in a library or plugin, unless the library or
>> > plugin imposes a similar requirement on its callers.
>> > """
>
> With that change it seems quite robust here.  I"ll push that in a
> short while unless I hear an objection from you.

Hmm it wasn't working yesterday with that change and something like
"sudo lxc-test-concurrent -j 20 -i 10" was getting stuck as soon as it
starts. it could be my system, and if it's working fine over there
then please feel free to push it.

> thanks,
> -serge
> _______________________________________________
> lxc-devel mailing list
> lxc-devel at lists.linuxcontainers.org
> http://lists.linuxcontainers.org/listinfo/lxc-devel

-- 
S.Çağlar Onur <caglar at 10ur.org>