New CentOS 7 container "connect: Network is unreachable"

[el_pedr0]

Hi,

I'm having major network difficulties with a fresh container install of CentOS 7 using the standard template and would welcome any help to resolve it.

I can connect via pct from the host, but I can't ping from the container, the response is:

[root@test ~]# ping 8.8.8.8
connect: Network is unreachable

If I try and ping the guest from the host, I get " Destination Host Unreachable"

In the guest, when I try:

[root@test ~]# systemctl restart network
Failed to get D-Bus connection: No such file or directory

Guest network config:

[root@test ~]# vi /etc/sysconfig/network-scripts/ifcfg-eth0
DEVICE=eth0
ONBOOT=yes
BOOTPROTO=static
IPADDR=192.168.1.19
NETMASK=255.255.255.0
GATEWAY=192.168.1.254

In the syslogs of the Proxmox host in the GUI, I see the following when I try and start the container:

Oct 07 22:19:37 toast.home.pmer.tk pvedaemon[2669]: <root@pam> starting task UPID:toast:0000245E:00003B56:615F6469:vzstart:119:root@pam:
Oct 07 22:19:37 toast.home.pmer.tk systemd[1]: Started PVE LXC Container: 119.
Oct 07 22:19:37 toast.home.pmer.tk pve-lxc-hook-lxc-pre-start[9387]: CT 119 does not support running in a pure cgroupv2 environment
Oct 07 22:19:37 toast.home.pmer.tk audit[9397]: AVC apparmor="STATUS" operation="profile_load" profile="/usr/bin/lxc-start" name="lxc-119_</var/lib/lxc>" pid=9397 comm="apparmor_parser"
Oct 07 22:19:37 toast.home.pmer.tk kernel: audit: type=1400 audit(1633641577.874:1813): apparmor="STATUS" operation="profile_load" profile="/usr/bin/lxc-start" name="lxc-119_</var/lib/lxc>" pid=9397 comm="apparmor_parser"
Oct 07 22:19:37 toast.home.pmer.tk systemd-udevd[9419]: ethtool: autonegotiation is unset or enabled, the speed and duplex are not writable.
Oct 07 22:19:37 toast.home.pmer.tk systemd-udevd[9419]: Using default interface naming scheme 'v247'.
Oct 07 22:19:38 toast.home.pmer.tk kernel: vmbr0: port 6(veth119i0) entered blocking state
Oct 07 22:19:38 toast.home.pmer.tk kernel: vmbr0: port 6(veth119i0) entered disabled state
Oct 07 22:19:38 toast.home.pmer.tk kernel: device veth119i0 entered promiscuous mode
Oct 07 22:19:38 toast.home.pmer.tk kernel: eth0: renamed from vethESajJU
Oct 07 22:19:38 toast.home.pmer.tk pvedaemon[2669]: <root@pam> end task UPID:toast:0000245E:00003B56:615F6469:vzstart:119:root@pam: WARNINGS: 1
Oct 07 22:19:38 toast.home.pmer.tk kernel: cgroup2: Unknown parameter 'mode'

When I double click to open a console on from the GUI, I am shown a black screen with 'Connected' at the top, but no log-in prompt. I can type characters into the console but they have no effect and I never get the console prompt that indicates a new line

Furthermore, I can't reboot or shutdown the guest - either from GUI or within the guest commandline:

[root@test ~]# reboot -h now
Failed to talk to init daemon.

[root@test ~]# shutdown -h now
Failed to talk to init daemon.

I'm running proxmox 7.0-11

Any help much appreciated.

 

[el_pedr0]

Hmm. Getting very similar behaviour with a new CentOS 8 container.

In the past I have created both CentOS 7 and CentOS 8 containers and haven't had network or dbus issues. Though I can't be sure of what version of Proxmox that was with - it was a few weeks back.

 

[t.lamprecht]

Hi,

I'm having major network difficulties with a fresh container install of CentOS 7 using the standard template and would welcome any help to resolve it.

CentOS 7 runs quite an old software stack and its's systemd is rather old and cannot cope with the cgroup2 unified mode Proxmox VE 7.x defaults too. That's why you should get a warning like the following in the CT create task log:

WARN: old systemd (< v232) detected, container won't run in a pure cgroupv2 environment! Please see documentation -> container -> cgroup version.

So the CT userland just mostly does not run as the old systemd version (the service manager) cannot run.

See: https://pve.proxmox.com/pve-docs/chapter-pct.html#pct_cgroup_compat