centos:7 container failed to start with /usr/sbin/init after update 12/16/2015

[centos]

centos:7 container failed to start with /usr/sbin/init after update 12/16/2015

[jperrin]

I don't understand why you're attempting to start systemd as an interactive container. What is the use case for doing this?

[nhlong]

Hi @jperrin, thanks for your comment
I just run it interactively to check the boot log and do some tests. In reality I run it as a daemon container.
In the old version it started well, but after the update there are some errors during boot as I mentioned above, the remount-fs and network service failed to start

systemctl status network
● network.service - LSB: Bring up/down networking
   Loaded: loaded (/etc/rc.d/init.d/network)
   Active: failed (Result: exit-code) since Thu 2015-12-17 03:33:34 UTC; 12min ago
     Docs: man:systemd-sysv-generator(8)

Dec 17 03:33:33 5aff87343190 systemd[1]: Starting LSB: Bring up/down networking...
Dec 17 03:33:34 5aff87343190 network[61]: Bringing up loopback interface:  [  OK  ]
Dec 17 03:33:34 5aff87343190 network[61]: Bringing up interface ens3:  ERROR    : [/etc/sysconfig/network-scripts/ifup-eth] Device ens3 does not seem to be present, delaying initialization.
Dec 17 03:33:34 5aff87343190 network[61]: [FAILED]
Dec 17 03:33:34 5aff87343190 systemd[1]: network.service: control process exited, code=exited status=1
Dec 17 03:33:34 5aff87343190 systemd[1]: Failed to start LSB: Bring up/down networking.
Dec 17 03:33:34 5aff87343190 systemd[1]: Unit network.service entered failed state.
Dec 17 03:33:34 5aff87343190 systemd[1]: network.service failed.

For network service issue, there is a redundant ifcfg-ens3, when removing this file network service can start well. However for remount-fs I cant find solution

In my Dockerfile, I has installed some packages

RUN yum install -y epel-release;
RUN yum -y update && yum clean all \
    && yum install -y --enablerepo=cr nginx \
    && yum install -y gettext pwgen initscripts sudo git ntp rsyslog autofs iptables iptables-services firewalld crond net-tools openssh-clients abrt-cli traceroute vim bind-utils openssh-server;

[jperrin]

If you're running the container based on 7.2.1511, there's no point to using the cr repo. It simply contained packages for 7.2.1511 before it was officially released. As I understand it, systemd still requires --privileged when running inside a container.

The "old" container used systemd-container, which was a reduced-functionality package, and it's been deprecated upstream. This is why it was replaced with systemd. If you feel this is a systemd issue, I would encourage you to file a bug at bugzilla.redhat.com against rhel 7.2

[jperrin]

The package list you're installing seems to imply that you're using this as fake virt, not generally what's envisioned with the container concept.

[nhlong]

Hi @jperrin, I tried with --privileged before but the boot process throwed more errors, for the old image it worked well.
--The package list you're installing seems to imply that you're using this as fake virt--
Yes, my app has run on VM before, Im trying to migrate it to docker container. Since it is started when booting system by systemd and requires some other facilities so I need to install that package list.
Thanks for your support.