On 03/13/2013 08:19 PM, Stéphane Graber wrote:
> On 03/13/2013 08:14 PM, Alexander Vladimirov wrote:
>> I am unable to reproduce it with my containers, can you show config
>> used to test?
>
> http://paste.ubuntu.com/5612341/
>
> As far as we can tell so far, the code looks fine but something is
On 03/13/2013 08:14 PM, Alexander Vladimirov wrote:
> I am unable to reproduce it with my containers, can you show config
> used to test?
http://paste.ubuntu.com/5612341/
As far as we can tell so far, the code looks fine but something is
somehow setting the value in the struct to -1. It's not any
I am unable to reproduce it with my containers, can you show config
used to test?
2013/3/14 Stéphane Graber :
> On 03/12/2013 09:37 AM, Serge Hallyn wrote:
>> Quoting Alexander Vladimirov (alexander.idkfa.vladimi...@gmail.com):
>>> I remember discussion about implementing proper way to shutdown
>>
I'm using Gentoo as my host OS (3.8.0 kernel), and the latest ebuild of
lxc is 0.8.0. Now it's possible (probable) that the udev that I'm
running is different than what is running on Ubuntu, but the behavior on
my system is that no devices are created. And yes, I've opened the
capabilities t
On 03/12/2013 09:37 AM, Serge Hallyn wrote:
> Quoting Alexander Vladimirov (alexander.idkfa.vladimi...@gmail.com):
>> I remember discussion about implementing proper way to shutdown
>> guests using different signals, so here's a patch proposal.
>> It allows to use specific signal numbers to shutdow
Thanks for fixing this up Serge, sorry about the regression!
On Wed, 13 Mar 2013 10:33:00 -0500
Serge Hallyn wrote:
> Otherwise containers fail to start even if they aren't trying to map
> ids.
>
> Also don't allocate buf unless we need to.
>
> Reported-by: Alexander Vladimirov
> Signed-off-b
Hello,
I'm going to start testing and prepare the pull request for rc1 in an
hour or so.
I think I've gone through any remaining patch on the mailing-list but
would appreciate if you could all check that the bits you care about
have been merged into staging.
If I forgot anything, please let me k
On 03/13/2013 11:33 AM, Serge Hallyn wrote:
> Otherwise containers fail to start even if they aren't trying to map
> ids.
>
> Also don't allocate buf unless we need to.
>
> Reported-by: Alexander Vladimirov
> Signed-off-by: Serge Hallyn
Looks good to me.
Acked-by: Stéphane Graber
> ---
> s
Otherwise containers fail to start even if they aren't trying to map
ids.
Also don't allocate buf unless we need to.
Reported-by: Alexander Vladimirov
Signed-off-by: Serge Hallyn
---
src/lxc/conf.c | 23 ++-
1 file changed, 14 insertions(+), 9 deletions(-)
diff --git a/src
Quoting Alexander Vladimirov (alexander.idkfa.vladimi...@gmail.com):
> With recent commits this code path prevents containers from starting
> on vanilla kernels, because there's no /proc//{u,g}id_map support
> there.
> If it was intended to make user namespaces support mandatory, then
> disregard t
With recent commits this code path prevents containers from starting
on vanilla kernels, because there's no /proc//{u,g}id_map support
there.
If it was intended to make user namespaces support mandatory, then
disregard that patch please.
2013/3/13 Serge Hallyn :
> Quoting Alexander Vladimirov (ale
Quoting Alexander Vladimirov (alexander.idkfa.vladimi...@gmail.com):
>
> Signed-off-by: Alexander Vladimirov
Acked-by: Serge E. Hallyn
> ---
> templates/lxc-archlinux.in | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/templates/lxc-archlinux.in b/templates/lxc-arc
Quoting Alexander Vladimirov (alexander.idkfa.vladimi...@gmail.com):
>
> Signed-off-by: Alexander Vladimirov
Nothing wrong with this, but why do you want this? It will save you
a 4096k allocation (which we could optimize away so it only happens
if needed), and one function call, but that is all
Signed-off-by: Alexander Vladimirov
---
src/lxc/start.c | 8 +---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/src/lxc/start.c b/src/lxc/start.c
index 3fdfa04..05646c2 100644
--- a/src/lxc/start.c
+++ b/src/lxc/start.c
@@ -823,9 +823,11 @@ int lxc_spawn(struct lxc_handler *h
Signed-off-by: Alexander Vladimirov
---
templates/lxc-archlinux.in | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/templates/lxc-archlinux.in b/templates/lxc-archlinux.in
index a542366..75c4d17 100644
--- a/templates/lxc-archlinux.in
+++ b/templates/lxc-archlinux.in
@@ -12
Ok, I'll apply your first version from Mar 11 and push a patch on
top of that to default it to 1. Thanks.
-serge
Quoting Alexander Vladimirov (alexander.idkfa.vladimi...@gmail.com):
> I do agree as well, and prefer to stick with first variant, but I was
> not sure what would be appropriate place
Quoting Stéphane Graber (stgra...@ubuntu.com):
...
> The short and rather usual reply to this is that we know we'll need a
> device namespace at some point. Exactly how it'll work is yet unknown
Yup.
> and it's never been high enough priority that anybody really worked on it.
>
> However I'm a b
I do agree as well, and prefer to stick with first variant, but I was
not sure what would be appropriate place to initialize default value
for that variable
2013/3/13 Serge Hallyn :
> Alexander it's up to you - I do agree with Michael, but I don't want
> any more of anyone's time spent on this, so
Quoting Alexander Vladimirov (alexander.idkfa.vladimi...@gmail.com):
> well, this is kind of workaround for workaround, so I guess this
> option won't stay there forever anyway
Right, hopefully we'll get namespaced /dev/kmsg soon :)
-serge
Alexander it's up to you - I do agree with Michael, but I don't want
any more of anyone's time spent on this, so if you want to stick with
what you sent, just reply with 'ack' and I'll ack and push this
patch.
thanks,
-serge
Quoting Alexander Vladimirov (alexander.idkfa.vladimi...@gmail.com):
> w
13.03.2013 10:05, Alexander Vladimirov wrote:
> Disables /dev/kmsg symlinking if set to 1.
Maybe it is better to avoid negative setting?
Like,
lxc.link_kmsg = 1
(and make it the default) ?
Thanks,
/mjt
--
Everyone ha
well, this is kind of workaround for workaround, so I guess this
option won't stay there forever anyway
2013/3/13 Michael Tokarev :
> 13.03.2013 10:05, Alexander Vladimirov wrote:
>> Disables /dev/kmsg symlinking if set to 1.
>
> Maybe it is better to avoid negative setting?
> Like,
>
> lxc.link_
22 matches
Mail list logo
