[Users] Debian 8 packages for OpenVZ as a HN?

[Jonathan Ballet]

Hi Kir,

On 05/12/2015 08:46 PM, Kir Kolyshkin wrote:
> On 05/12/2015 02:00 AM, Jonathan Ballet wrote:
>>
>>
>> Hi,
>>
>> On 05/10/2015 05:05 PM, Pavel Odintsov wrote:
>>> Well, well. But why my 2.6.32 kernel become broken after change
>>> Wheezy's init system to systemd? Standard 3.2 kernel from Debian works
>>> perfectly with systemd.
>>>
>>> If this problems is not related with kdbus but it's still it broke
>>> ability to run OpenVZ kernel on modern distros.
>>>
>>> If somebody will fix this issue I will be very pleased.
>>
>> I tried booting one of the latest OpenVZ release on Debian Jessie with 
>> systemd, and I got the error:
>>
>>   Failed to mount /sys/fs/cgroup: No such file or directory.
>>
>> This is related to this bug :
>>
>>   https://bugzilla.redhat.com/show_bug.cgi?id=628004
>>
>> Basically (AFAIU), the OpenVZ kernel lacks support for mounting 
>> cgroups at the required place, which make systemd choke when starting 
>> up in the early phase.
>> As mentioned in the bug report above, applying the patch from [1] 
>> might be enough to "fix" the problem, but I haven't tested this one 
>> and I'm not sure about the consequences.
> 
> I'm afraid you are wrong here, at least It's not the case for me --
> with 2.6.32-042stab108.1 kernel and CentOS 7 OS template.
> 
> See:
> 
> [root at tpad-ovz1 ~]# vzctl enter 101
> entered into CT 101
> [root at efgh /]# ls -l /sys/fs
> total 0
> drwxr-xr-x 10 root root 200 Apr 28 01:04 cgroup
> [root at efgh /]# ls -l /sys/fs/cgroup/
> total 0
> drwxr-xr-x 2 root root  0 Apr 28 01:04 blkio
> drwxr-xr-x 2 root root 40 Apr 28 01:04 cpu,cpuacct
> drwxr-xr-x 2 root root 40 Apr 28 01:04 cpuset
> drwxr-xr-x 2 root root 40 Apr 28 01:04 freezer
> drwxr-xr-x 2 root root  0 Apr 28 01:04 memory
> drwxr-xr-x 2 root root 40 Apr 28 01:04 net_cls,net_prio
> drwxr-xr-x 2 root root 40 Apr 28 01:04 perf_event
> drwxr-xr-x 4 root root  0 Apr 28 01:04 systemd
> 
> [root at efgh /]# mount | grep cgroup
> tmpfs on /sys/fs/cgroup type tmpfs (rw,nosuid,nodev,noexec,mode=755)
> cgroup on /sys/fs/cgroup/systemd type cgroup 
> (rw,nosuid,nodev,noexec,relatime,release_agent=/usr/lib/systemd/systemd-cgroups-agent,name=systemd) 
> 
> cgroup on /sys/fs/cgroup/blkio type cgroup 
> (rw,nosuid,nodev,noexec,relatime,blkio,name=beancounter)
> cgroup on /sys/fs/cgroup/memory type cgroup 
> (rw,nosuid,nodev,noexec,relatime,memory)
> 
> [root at efgh /]# logout
> exited from CT 101
> 
> [root at tpad-ovz1 ~]# vzlist -o ostemplate 101
> OSTEMPLATE
> centos-7-x86_64-minimal
> 
> [root at tpad-ovz1 ~]# uname -a
> Linux tpad-ovz1 2.6.32-042stab108.1 #1 SMP Thu Apr 23 19:17:11 MSK 2015 
> x86_64 x86_64 x86_64 GNU/Linux

I'm not sure what's your point here. I was talking about running
systemd on the HN, not in a container. Your kernel seem to be
slightly newer than the version I'm running, so maybe things have
changed in between.
*This* configuration (OpenVZ kernel + systemd on the HN) currently
doesn't work because of the lack of /sys/fs/cgroup on the HN which
prevents systemd from starting.

  root@:~ # uname -a
  Linux hn-laus4 2.6.32-openvz-042stab106.6-amd64 #1 SMP Mon Apr 20 14:57:06 MSK 2015 x86_64 GNU/Linux
  root@:~ # cat /proc/vz/version
  042stab106_6
  root@:~ # ls -l /sys/fs/cgroup
  ls: cannot access /sys/fs/cgroup: No such file or directory

Maybe this has changed with kernel version 2.6.32-042stab108?

I'm not sure how you get access to /sys/fs/cgroup in the container
though, I would have believed this was mounted by the HN, but I
haven't tried running a newer version (ie. Jessie, with systemd)
in a container yet.

 Jonathan