2 PM, jjs - mainphrame
> wrote:
> > > I'm running both lxc and lxd containers on the same box. There is no
> > > conflict, as the commands and data paths are different.
> > >
> > > Jake
> > >
> > > On Mon, Apr 11, 2016 at 5:12 PM, Bos
On 12 April 2016 at 02:32, jjs - mainphrame wrote:
> I'm running both lxc and lxd containers on the same box. There is no
> conflict, as the commands and data paths are different.
>
Excellent news!
Are you running Ubuntu, some other distro or custom rolled lx* installs?
tnx,
b.
___
Is it possible?
Or is running LXC on host and LXD as one of LXC containers more inclined to
work out of the box?
(just fishing for info, haven't tried LXD out yet)
Tnx,
b.
___
lxc-users mailing list
lxc-users@lists.linuxcontainers.org
http://lists.linux
Dear Akshay,
I do agree with you and find this behaviour a bit annoying, yet I believe
"patches welcome" response will follow shortly :)
On a more serious note:
As I skimmed over LXC code a while ago, it seems LXC bails out on first
error that occurs. This means that implementing your suggestion
I think 770 permission on /home/gmc/.local/share/lxc/xxx or somewhere above
are the cause of it.
Usually when I create container and uidmapshift it (rootfs), I have to
chmod 755 /var/lib/lxc/container-name before it can start. This chmod step
is not needed for privileged containers.
I think this
container, i
> have to get rid of it to start the Container again. I use
>
> ip link set dev $DEV name away.$RANDOM
>
> to archive this
>
> Guido
>
>
>
>
> On 19.02.2016 17:08, Bostjan Skufca wrote:
> > Does it delete both interfaces, the on
Does it delete both interfaces, the one in host's namespace which is added
to bridge (vethX) and the one in container's namespace (appears as ethX in
there) too?
b.
On 19 February 2016 at 11:58, Fajar A. Nugraha wrote:
> On Fri, Feb 19, 2016 at 5:41 PM, Bostjan Skufca wrot
I had similar experience using CentOS 7 as host.
I did not investigate it yet, but for some reason veth network interfaces
were not being destroyed when containers stopped, and this could only be
seen with "brctl show" as more interfaces were attached to a bridge than
there were containers running
ib/python3.5/lib-dynload/_bz2.cpython-35m-x86_64-linux-gnu.so
> ",
> O_RDONLY|O_CLOEXEC) = 3
> open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3
> open("/usr/lib/libbz2.so.1.0", O_RDONLY|O_CLOEXEC) = 3
> open("/usr/lib/python3.5/__pycache__/lzma.cpython
strace command is your friend.
b.
On 21 January 2016 at 12:12, Viktor Trojanovic wrote:
>
>
> On 20.01.2016 23:50, Fajar A. Nugraha wrote:
>
> On Thu, Jan 21, 2016 at 5:49 AM, Fajar A. Nugraha <
> l...@fajar.net> wrote:
>
>> On Thu, Jan 21, 2016 at 5:23 AM, Viktor Trojanovic
>> wrote:
>>
>>>
I agree lxc-top is somewhat confusing. For example, all columns except Mem
show totals, while Mem shows current memory usage.
I believe lxc-top needs some work to become actually usable. Data it
displays should be in one of the following modes:
- current value, like top
- total divided by containe
ichael wrote:
> What would it be the right tar parameters to compress and decompress all
> the rootfs, including devices and special files?
>
> On Fri, Dec 4, 2015 at 11:58 AM, Bostjan Skufca wrote:
>
>> Depends if you need consistent copy and how much downtime you can
>&g
ichael wrote:
> What would it be the right tar parameters to compress and decompress all
> the rootfs, including devices and special files?
>
> On Fri, Dec 4, 2015 at 11:58 AM, Bostjan Skufca wrote:
>
>> Depends if you need consistent copy and how much downtime you can
>&g
Depends if you need consistent copy and how much downtime you can tolerate.
If inconsistent copy is enough, then you can run rsync over storage of
running container (on host, not in container) and be done with it.
Rsync:
I find rsync useful and fast, providing that:
- whole container filesystem m
I recently asked a similar questions (did not mention rdb though that was
exactly what I had in my mind:), maybe you will find Stephane's answer
informative:
https://lists.linuxcontainers.org/pipermail/lxc-users/2015-November/010414.html
b.
On 25 November 2015 at 16:40, Dilvan Moreira wrote:
>
... or the one from Georg:
https://lists.linuxcontainers.org/pipermail/lxc-users/2015-November/010417.html
On 25 November 2015 at 18:48, Bostjan Skufca wrote:
> I recently asked a similar questions (did not mention rdb though that was
> exactly what I had in my mind:), maybe you wil
t;>> web50
>>> web51
>>> webWithOtherService
>>>
>>> $lxc start [pattern], depends on the pattern I can several containers
>>> with common names.
>>>
>>> Even though, I find it useful to start and stop container, using patterns
>
Speaking off the top of my head, but does not lxc-stop send sigterm signal
to all processes in a container, and then waits for configurable amount of
time (default=60s), before sending sigkill?
If your processes expect to be shut down by init scripts that need to do
some additional tasks, well, th
+1
Luis: how do you see this implemented, as simple shell-like globbing, or
full regex support?
b.
On 13 November 2015 at 14:29, Luis Michael Ibarra
wrote:
>
> Hi all,
>
> I think this is a minor feature, but it would be nice to have globbing
> support for lxc [command] parameters. I do know
> Looks like lxc-clone should copy the config file at the very end, after
> rootfs.
+1
Use cp analogy for example (with behaviour that most users got used to
expect in the last few decades):
cp -pR dir1 dir2
# (cp gets interrupted for some reason, ctrl+c or whatever, ok, I need
to start again)
LXC/LXD? How did you move the containers? Did you use rsync without
--numeric-ids maybe?
b.
On 10 November 2015 at 10:37, Jamie Brown wrote:
> Correction; the user was changed to “ubuntu” but the group was changed to
> “admin”. Though I’m guessing it is more to do with matching of the
> underl
Hello,
if one wants to have container host cluster with HA feature (restarting
containers on non-failed hosts), is this something that is planned for LXD,
or is going the OpenStack route the way to go for the foreseeable future?
Am I missing some other already existing product (Proxmox I know of)?
Hello Tycho
On 9 November 2015 at 21:30, Tycho Andersen
wrote:
> Hello Boštjan,
>
> On Mon, Nov 09, 2015 at 06:47:42PM +0100, Boštjan Škufca @ Teon.si wrote:
> > Containers start, but this is what I am getting:
> > lxc-start: utils.c: setproctitle: 1461 Invalid argument - setting cmdline
> > fai
Depends on what your requirements for "production" are. Live migration? I
guess not. Environment isolation for more-or-less trusted containers? Yes,
using it here for quite a while (since 1.0.6 - lxc, not lxd), if possible
as unprivileged containers, as it removes A LOT of attack surface for host.
This seems like NAT issue on the host. Did you do "iptables -t nat -I -o
lxcbrX -o phyY -j MASQUERADE"?
b.
On 14 October 2015 at 11:25, wrote:
> I use raw socket to send an ip packet and fill the source ip 5.5.5.5. I
> use tcpdump to capture it:
>
> * 15:01:29.385061 IP 5.5.5.5 > 224.0.0.18
Hi Teemu,
does this tool have a name?
IMHO it is missing -r switch that is present in uidmapshift.c (info about
existing uids, gids).
Additionally, uidmapshift can limit its scope of operation, and not just
directory-wise, but also by id-range. For example, if you run uidmapshift
on a directory
Mark, you are probably talking about lxD, not lxC, right?
b.
On 6 October 2015 at 07:36, Mark Constable wrote:
> lxc v0.19 on Ubuntu 15.10 host.
>
> ~ lxc launch wily abc
> Creating abc done.
> Starting abc done.
>
> ~ lxc launch wily abc.lxc
> Creating abc.lxc error: Invalid container name
>
On 15 September 2015 at 19:46, Serge Hallyn wrote:
> It sounds like it may be worthwhile. The patch shouldn't be huge, so
> I think it's worth creating the patch and sending it to the list. Do
> make sure to give a detailed description of how you'll use it. (Don't
> assume I'll remember :)
Tnx
process that is starting the
lxc-start is the way currently, but awkward.
b.
On 14 September 2015 at 18:45, Serge Hallyn wrote:
> Which limits would you be intending to set?
>
> Quoting Bostjan Skufca (bost...@a2o.si):
>> (there was no response on this subject)
>>
>> I check
11:18, Bostjan Skufca wrote:
> Hi all,
>
> is there any elegant way for setting unpriv. container ulimits, other
> than setting it in current shell and starting container from that
> shell (and scripting all this together)?
>
> Hooks do not work, as they are executed a
Hi Peter,
since you mentioned you are using bridged interfaces, is my assumption
correct that your containers's network connection is joined directly
to this bridge and containers talk to the world direcly (L2) and not
via routed (L3) network over host OS?
Did you try using routed setup (using bo
Hi all,
is there any elegant way for setting unpriv. container ulimits, other
than setting it in current shell and starting container from that
shell (and scripting all this together)?
Hooks do not work, as they are executed as subprocesses.
b.
___
lxc
I did not do what you describe exactly, but went through adaptation of
regular linux to support container-based version.
I short, what you need to take care of, is:
- /dev - copy from one of existing containers, replace. You need very few
devices in container.
- filesystems /proc, dev and sys
- ud
Hi Siva,
how does your output of command "lxc-checkconfig" look like?
b.
On 1 August 2015 at 00:11, SIVA SUBRAMANIAN.P wrote:
> Hi,
> I'm trying to generate a image with lxc 1.1.2 on 3.14 kernel, but not able
> to boot it. Is there any hard dependency I need to account in?
>
> Regards,
> Siv
On 15 July 2015 at 08:29, Christoph Mathys wrote:
> On Wed, Jul 15, 2015 at 4:08 AM, Bostjan Skufca wrote:
> >
> >
> > Does anyone know where that container variable gets its value from?
> >
>
> src/lxc/start.c in do_start() contains:
>
> if (putenv("
On 14 July 2015 at 22:33, Tycho Andersen
wrote:
> On Mon, Jul 13, 2015 at 10:53:10AM +0200, Christoph Mathys wrote:
> > Is there an unhacky way of knowing if a script runs inside a
> > container? In my case, a sysV initscript that tries to load some
> > kernel modules needs to know if it runs ins
Is systemd now supported as LXC guest's init system?
Anyhow, I would guess systemd creates ramdisk for /run (and /var/run)
which, by some funky interaction ends up non-writeable. I had (somewhat)
similar systemd issue on another project recently, where systemd created
/dev/log socket preemptively,
Hi Serge,
is there any standard implementation for starting user-unprivileged
containers at boot? I am not talking about containers which are uidmapped
(and started) by root to be unprivileged. I mean containers which are
created by unprivileged users in their home dirs.
Tnx for info,
b.
On 3 A
On 24 March 2015 at 15:34, Tycho Andersen
wrote:
> Hi,
>
> You'll also find that you need some other stuff to do live migration
> (not documented anywhere, unfortunately):
> ...
> 2. only non-uidmapped containers work (you can do this in lxd by just
>setting the uidmap for lxd to be 0-65536,
I guess without additional data (container network configuration, host
network configuration, and ifconfig and route output to check if specified
configuration actually matches the current state of your system), no one
will be able to help you.
b.
On 23 March 2015 at 16:41, Thouraya TH wrote:
Do you want your containers to use public IPs direcly, or do you want to
use local IPs and then forward whole traffic for certain IP towards
particular container?
In first case, just set lxc.network.link = br0 and configure public IP as
you have described it. Container uses the same GW as your hos
What is your host running?
b.
On 6 March 2015 at 22:25, CDR wrote:
> Downloading packages:
> mtr-0.85-7.el7.x86_64.rpm
> | 71 kB 00:00:00
> Running transaction check
> Running transaction test
> Transaction test succeeded
> Running transaction
> Installing :
> 2:mtr-0.85-7.el7.x86_64
> 1/1
LXC 1.1.0:
I am getting this error which I fail to fully understand ATM, as I did not
dig around code too much yet.
If I try to start a container with autodev = 1, it fails to start and
results in this error:
lxc-start: conf.c: setup_ttydir_console: 1525 Device or resource busy -
error unlinking /
On 6 March 2015 at 13:41, Bostjan Skufca wrote:
>
>
> On 6 March 2015 at 08:21, Fajar A. Nugraha wrote:
>
>> On Fri, Mar 6, 2015 at 8:40 AM, Bostjan Skufca wrote:
>> > How compatible is LXC 1.1.x with 1.0.7?
>>
> > I do not have problems with updati
On 5 March 2015 at 03:32, Fajar A. Nugraha wrote:
> On Thu, Mar 5, 2015 at 9:21 AM, Bostjan Skufca wrote:
> > Here are the steps to reproduce (lxc 1.0.7), assuming container is
> already
> > running:
>
> > ### Step 3: watch the shutdown messages in console on h
On 6 March 2015 at 08:21, Fajar A. Nugraha wrote:
> On Fri, Mar 6, 2015 at 8:40 AM, Bostjan Skufca wrote:
> > How compatible is LXC 1.1.x with 1.0.7?
>
> Did you try it?
>
Not yet, will do in the following days (machine has just been acquired for
this).
> > I d
On 5 March 2015 at 03:32, Fajar A. Nugraha wrote:
>
> Works for me. Try upgrading lxc.
>
How compatible is LXC 1.1.x with 1.0.7?
I do not have problems with updating config files, but are there any side
effects to be expected (like containers not starting for no apparent
reason, etc)?
Is 1.1.1
On 5 March 2015 at 09:23, Jäkel, Guido wrote:
> Dear Bostjan,
>
> I already told my 5ct on this many month ago, but now -- with LXC 1.x --
> it might be the time to discuss about it, again.
>
>
> IMHO one want to describe dependencies and it's up to the computer to
> derive any order from that. I
Hi there,
if container is rebooted while someone is attached to its primary console
(/dev/lxc/console), then than console becomes "unreachable" until container
is shut down and started again. Consoles /dev/lxc/ttyX work as expected all
the time.
Here are the steps to reproduce (lxc 1.0.7), assumi
Hi there,
I would like to open a discussion about container ordering regarding to
lxc.start.order and lxc-autostart operations.
Currently, let's presume that pull request
https://github.com/lxc/lxc/pull/461 is merged and that containers start in
ascending lxc.start.order fashion.
This makes sens
On 4 March 2015 at 14:48, Jean-Gabriel Gill-Couture <
jeangabriel...@gmail.com> wrote:
>
> Is it possible that the order is respected but the first containers are
> just slower to start up?
>
Nope, just checked it again, with longer delays to be able to watch what is
going on. Containers with orde
I understood from man pages that lxc.start.order setting should cause
containers to start in ascending order (lower the value, earlier the
startup). It turns out that with 1.0.7 this acts more like priority - the
higher the value, the sooner container starts.
Is anyone else experiencing this with
If I want to do something like that (i.e. when bootstrapping custom linux
system as LXC guest), I just chroot into newly-created container root and
do the "apt-get install ...".
This method uses host's network connection, but it requires container's
/etc/resolv.conf correctly configured. Also, if
Hope? :)
To get the actual value, one should probably explore uid_t type definition
if you want to skip reading sources of the shadow suite. Online search
returns this nice and clickable page:
http://lxr.free-electrons.com/ident?i=uid_t
This leads to "unsigned int" as final definition. On 64-bit
Hi all,
I was just debugging some IPv6 connectivity problems and realised that
container interfaces do not inherit host's default settings about IPv6
autoconfiguration.
More precisely I am talking about these two parameters:
- /proc/sys/net/ipv6/conf/*/accept_ra
- /proc/sys/net/ipv6/conf/*/autoco
This is not lxc-specific, but James Bottomley gave a insightful talk about
containers vs full VMs:
http://www.youtube.com/watch?v=p-x9wC94E38
Charts around 14min gives nice general (ballpark) comparison figures.
b.
On 16 October 2014 08:33, Ritesh Ranjan wrote:
> Any docs/links that some one
Hi,
I am using LXC quite successfully and most of my containers are
already ported to unprivileged mode of operation (started from root,
but uidmapped).
One thing that is getting in the way are CIFS mounts.
If I mount it either via pre-start hook or via lxc.mount.entry config
variable, when cont
57 matches
Mail list logo