After Kernel-Upgrade many containers not start

[Virtualizer]

I have upgrade today to the new kernel and massive problems begun! Before was all ok!

uname -a
Linux pm20-host 4.13.13-5-pve #1 SMP PVE 4.13.13-38 (Fri, 26 Jan 2018 10:47:09 +0100) x86_64 GNU/Linux

A lot of containers start normal and others cant start and I see in log file:

Jan 27 09:18:50 pm20-host systemd-udevd[7203]: Could not generate persistent MAC address for vethEA453P: No such file or directory
Jan 27 09:18:50 pm20-host systemd-udevd[7274]: Could not generate persistent MAC address for fwbr1226i0: No such file or directory
Jan 27 09:18:50 pm20-host systemd-udevd[7283]: Could not generate persistent MAC address for fwpr1226p0: No such file or directory
Jan 27 09:18:50 pm20-host systemd-udevd[7284]: Could not generate persistent MAC address for fwln1226i0: No such file or directory
Jan 27 09:18:50 pm20-host lxc-start[7154]: lxc-start: 1226: lxccontainer.c: wait_on_daemonized_start: 760 Received container state "STOPPING" instead of "RUNNING"
Jan 27 09:18:50 pm20-host lxc-start[7154]: lxc-start: 1226: tools/lxc_start.c: main: 371 The container failed to start.
Jan 27 09:18:50 pm20-host lxc-start[7154]: lxc-start: 1226: tools/lxc_start.c: main: 373 To get more details, run the container in foreground mode.
Jan 27 09:18:50 pm20-host lxc-start[7154]: lxc-start: 1226: tools/lxc_start.c: main: 375 Additional information can be obtained by setting the --logfile and --logpriority options.
Jan 27 09:18:50 pm20-host systemd[1]: pve-container@1226.service: Control process exited, code=exited status=1
Jan 27 09:18:50 pm20-host systemd[1]: pve-container@1226.service: Killing process 7156 (lxc-start) with signal SIGKILL.
Jan 27 09:18:50 pm20-host systemd[1]: pve-container@1226.service: Killing process 7390 (sh) with signal SIGKILL.
Jan 27 09:18:50 pm20-host systemd[1]: Failed to start PVE LXC Container: 1226.
Jan 27 09:18:50 pm20-host systemd[1]: pve-container@1226.service: Unit entered failed state.
Jan 27 09:18:50 pm20-host systemd[1]: pve-container@1226.service: Failed with result 'exit-code'.
Jan 27 09:18:50 pm20-host pve-guests[7149]: command 'systemctl start pve-container@1226' failed: exit code 1
Jan 27 09:18:50 pm20-host pve-guests[7403]: starting CT 1235: UPIDm20-host:00001CEB:0000857D:5A6C35EA:vzstart:1235:root@pam:
Jan 27 09:18:50 pm20-host systemd[1]: Starting PVE LXC Container: 1235...
Jan 27 09:18:51 pm20-host systemd-udevd[8196]: Could not generate persistent MAC address for vethENT2FR: No such file or directory
Jan 27 09:18:51 pm20-host systemd-udevd[8338]: Could not generate persistent MAC address for fwbr1235i0: No such file or directory
Jan 27 09:18:51 pm20-host systemd-udevd[8350]: Could not generate persistent MAC address for fwpr1235p0: No such file or directory
Jan 27 09:18:51 pm20-host systemd-udevd[8351]: Could not generate persistent MAC address for fwln1235i0: No such file or directory
Jan 27 09:18:51 pm20-host lxc-start[7406]: lxc-start: 1235: lxccontainer.c: wait_on_daemonized_start: 760 Received container state "STOPPING" instead of "RUNNING"
Jan 27 09:18:51 pm20-host lxc-start[7406]: lxc-start: 1235: tools/lxc_start.c: main: 371 The container failed to start.
Jan 27 09:18:51 pm20-host lxc-start[7406]: lxc-start: 1235: tools/lxc_start.c: main: 373 To get more details, run the container in foreground mode.
Jan 27 09:18:51 pm20-host lxc-start[7406]: lxc-start: 1235: tools/lxc_start.c: main: 375 Additional information can be obtained by setting the --logfile and --logpriority options.
Jan 27 09:18:51 pm20-host systemd[1]: pve-container@1235.service: Control process exited, code=exited status=1
Jan 27 09:18:51 pm20-host systemd[1]: pve-container@1235.service: Killing process 7408 (lxc-start) with signal SIGKILL.
Jan 27 09:18:51 pm20-host systemd[1]: pve-container@1235.service: Killing process 8451 (sh) with signal SIGKILL.
Jan 27 09:18:51 pm20-host systemd[1]: Failed to start PVE LXC Container: 1235.
Jan 27 09:18:51 pm20-host systemd[1]: pve-container@1235.service: Unit entered failed state.
Jan 27 09:18:51 pm20-host systemd[1]: pve-container@1235.service: Failed with result 'exit-code'.
Jan 27 09:18:51 pm20-host pve-guests[7403]: command 'systemctl start pve-container@1235' failed: exit code 1

So, I like to change to the old kernel, but how can I doe? I have test in /etc/default/grub - change default to 1 - and update-grub, but this helps not!

Thanks

Virtualizer

 

[Virtualizer]

I have now startet in old kernel-mode and thats the same!

Kernel uname -a
Linux pm20-host 4.13.13-4-pve #1 SMP PVE 4.13.13-35 (Mon, 8 Jan 2018 10:26:58 +0100) x86_64 GNU/Linux

I search more - has somewhere an idea?

 

[Virtualizer]

So I found a new big horrible Proxmox Bug and a workarround:

From one to other versions in PVE 5.x Containers not start, when in the Network-Configuration is set a speed limit (rate limit of network interface)! Where the same problem get from one day to another, here is the workarround:

In the GUI where the container not start, edit Container -> Network -> Network-Interface and their delete the counter (to a blank field), then you get "unlimited". Do the same with 2nd or more interfaces. After this, you can start the Container!

I hope this helps others too and I will write a bug report!

Regards

Detlef

 

[andaris]

Thank you very much, Detlef! I had the same problem here...

 

[puertorico]

• same issue here.

 

[fabian]

you can downgrade iproute as an interim measure:

apt-get install --reinstall iproute2=4.10.0-1

updated libpve-common-perl or iproute2 packages should be available soon.