[SOLVED] DHCP issues with Debian 10 Container

[Stargate85]

Hi all,

I run a Intel Nuc running Proxmox 6 (Linux 5.0.21-1-pve #1 SMP PVE 5.0.21-2) at Home with about 6 Containers and 3 VMs.
I have the problem that Debian 10 LXC Container are stuck at boot for about 5-10 minutes. So if I check the console its just black.

If I login through the host with lxc-attach I see this in the syslog. Any idea? For all other VMS/Containers (except Debian 10) DHCP runs fine (IPv4 and v6):

Sep 19 19:18:33 nuc-iobroker ifup[71]: XMT: Forming Solicit, 3150 ms elapsed.
Sep 19 19:18:33 nuc-iobroker ifup[71]: XMT:  X-- IA_NA 44:17:2a:e1
Sep 19 19:18:33 nuc-iobroker ifup[71]: XMT:  | X-- Request renew in  +3600
Sep 19 19:18:33 nuc-iobroker ifup[71]: XMT:  | X-- Request rebind in +5400
Sep 19 19:18:33 nuc-iobroker ifup[71]: XMT: Solicit on eth0, interval 4190ms.
Sep 19 19:18:33 nuc-iobroker ifup[71]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:18:33 nuc-iobroker ifup[71]: RCV:  X-- Server ID: 00:03:00:01:98:9b:cb:50:64:ff
Sep 19 19:18:33 nuc-iobroker ifup[71]: PRC: Lease failed to satisfy.
Sep 19 19:18:33 nuc-iobroker dhclient[120]: XMT: Solicit on eth0, interval 4190ms.
Sep 19 19:18:33 nuc-iobroker dhclient[120]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:18:37 nuc-iobroker ifup[71]: XMT: Forming Solicit, 7340 ms elapsed.
Sep 19 19:18:37 nuc-iobroker ifup[71]: XMT:  X-- IA_NA 44:17:2a:e1
Sep 19 19:18:37 nuc-iobroker ifup[71]: XMT:  | X-- Request renew in  +3600
Sep 19 19:18:37 nuc-iobroker ifup[71]: XMT:  | X-- Request rebind in +5400
Sep 19 19:18:37 nuc-iobroker ifup[71]: XMT: Solicit on eth0, interval 8020ms.
Sep 19 19:18:37 nuc-iobroker ifup[71]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:18:37 nuc-iobroker ifup[71]: RCV:  X-- Server ID: 00:03:00:01:98:9b:cb:50:64:ff
Sep 19 19:18:37 nuc-iobroker ifup[71]: PRC: Lease failed to satisfy.
Sep 19 19:18:37 nuc-iobroker dhclient[120]: XMT: Solicit on eth0, interval 8020ms.
Sep 19 19:18:37 nuc-iobroker dhclient[120]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:18:45 nuc-iobroker ifup[71]: XMT: Forming Solicit, 15370 ms elapsed.
Sep 19 19:18:45 nuc-iobroker ifup[71]: XMT:  X-- IA_NA 44:17:2a:e1
Sep 19 19:18:45 nuc-iobroker ifup[71]: XMT:  | X-- Request renew in  +3600
Sep 19 19:18:45 nuc-iobroker ifup[71]: XMT:  | X-- Request rebind in +5400
Sep 19 19:18:45 nuc-iobroker ifup[71]: XMT: Solicit on eth0, interval 16420ms.
Sep 19 19:18:45 nuc-iobroker ifup[71]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:18:45 nuc-iobroker ifup[71]: RCV:  X-- Server ID: 00:03:00:01:98:9b:cb:50:64:ff
Sep 19 19:18:45 nuc-iobroker ifup[71]: PRC: Lease failed to satisfy.
Sep 19 19:18:45 nuc-iobroker dhclient[120]: XMT: Solicit on eth0, interval 16420ms.
Sep 19 19:18:45 nuc-iobroker dhclient[120]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:19:02 nuc-iobroker ifup[71]: XMT: Forming Solicit, 31810 ms elapsed.
Sep 19 19:19:02 nuc-iobroker ifup[71]: XMT:  X-- IA_NA 44:17:2a:e1
Sep 19 19:19:02 nuc-iobroker ifup[71]: XMT:  | X-- Request renew in  +3600
Sep 19 19:19:02 nuc-iobroker ifup[71]: XMT:  | X-- Request rebind in +5400
Sep 19 19:19:02 nuc-iobroker ifup[71]: XMT: Solicit on eth0, interval 32220ms.
Sep 19 19:19:02 nuc-iobroker ifup[71]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:19:02 nuc-iobroker ifup[71]: RCV:  X-- Server ID: 00:03:00:01:98:9b:cb:50:64:ff
Sep 19 19:19:02 nuc-iobroker ifup[71]: PRC: Lease failed to satisfy.
Sep 19 19:19:02 nuc-iobroker dhclient[120]: XMT: Solicit on eth0, interval 32220ms.
Sep 19 19:19:02 nuc-iobroker dhclient[120]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:19:34 nuc-iobroker ifup[71]: XMT: Forming Solicit, 64060 ms elapsed.
Sep 19 19:19:34 nuc-iobroker ifup[71]: XMT:  X-- IA_NA 44:17:2a:e1
Sep 19 19:19:34 nuc-iobroker ifup[71]: XMT:  | X-- Request renew in  +3600
Sep 19 19:19:34 nuc-iobroker ifup[71]: XMT:  | X-- Request rebind in +5400
Sep 19 19:19:34 nuc-iobroker ifup[71]: XMT: Solicit on eth0, interval 62440ms.
Sep 19 19:19:34 nuc-iobroker ifup[71]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:19:34 nuc-iobroker ifup[71]: RCV:  X-- Server ID: 00:03:00:01:98:9b:cb:50:64:ff
Sep 19 19:19:34 nuc-iobroker ifup[71]: PRC: Lease failed to satisfy.
Sep 19 19:19:34 nuc-iobroker dhclient[120]: XMT: Solicit on eth0, interval 62440ms.
Sep 19 19:19:34 nuc-iobroker dhclient[120]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:20:36 nuc-iobroker ifup[71]: XMT: Forming Solicit, 126560 ms elapsed.
Sep 19 19:20:36 nuc-iobroker ifup[71]: XMT:  X-- IA_NA 44:17:2a:e1
Sep 19 19:20:36 nuc-iobroker ifup[71]: XMT:  | X-- Request renew in  +3600
Sep 19 19:20:36 nuc-iobroker ifup[71]: XMT:  | X-- Request rebind in +5400
Sep 19 19:20:36 nuc-iobroker ifup[71]: XMT: Solicit on eth0, interval 118780ms.
Sep 19 19:20:36 nuc-iobroker dhclient[120]: XMT: Solicit on eth0, interval 118780ms.
Sep 19 19:20:36 nuc-iobroker ifup[71]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.
Sep 19 19:20:36 nuc-iobroker ifup[71]: RCV:  X-- Server ID: 00:03:00:01:98:9b:cb:50:64:ff
Sep 19 19:20:36 nuc-iobroker ifup[71]: PRC: Lease failed to satisfy.
Sep 19 19:20:36 nuc-iobroker dhclient[120]: RCV: Advertise message on eth0 from fe80::9a9b:cbff:fe50:64ff.

Thats the interfaces

auto lo
iface lo inet loopback

auto eth0
iface eth0 inet6 dhcp


iface eth0 inet dhcp

 

[Stargate85]

Just found the issue myself. After setting IPv6 to static (before DHCP) in the Network Options solved it.

 

[kappclark]

I have been having the same issue -- I create the Debian 10 container, with network settings of static for IPV4 and IPV6 (which appear to be default) .. start the machine...

Then I issue dhclient command, get an address, and do update and upgrade.

Then I reboot the machine from the cli with reboot command.

Now, the CT will not start -- when I goto console, it says

lxc-console: 745: tools/lxc_console.c: main: 152 745 is not running

I click on the Start button, and I get:

ob for pve-container@745.service failed because the control process exited with error code.
See "systemctl status pve-container@745.service" and "journalctl -xe" for details.
TASK ERROR: command 'systemctl start pve-container@745' failed: exit code 1

Should I edit interfaces file before reboot ?? FWIW, I have no such 'challenges' with the Debian 9 container ...

 

[kappclark]

FWIW - here are the settings for the container which made it work: I changed IPV6 to SLAAC. Rebbots right away now...the DHCP server is not configured for IPV6 addresses - should I worry about this in a small home network ??

 

[kappclark]

OK - sorry to belabor this -- after update - all good (dns, dig etc) ... then after reboot, cannot access.....

Appears that the update-reboot will cripple the Container .... I have confirmed this on 2 nodes..

Install ct with static defaults
dhclient (to get connectivity
reboot -----> NO PROBLEM
apt update apt upgrade
reboot -----> will not start

No such 'issue' with debian 9 CT ...

 

[kappclark]

OK - Found this comment on the very same issue, but cannot understand what file needs to be changed ... he claims the exact update issue, and provided a remedy

*FWIW - I do not have an enterprise subscription) ... get some error messages when I try to update the proxmox host ...

Here is his post

********************** Update ****************
Jan 8 - I have 2 Proxmox hosts, and applied updates to both. Once buster installed on host, the problem goes away --

 

[neiion]

Hi
I have noticed the same issue running debian 10 lxc . it was taking a long time to start up. to solve the issue i disabled ip6 fully in lxc . to disable do the following in the lxc console:
nano /etc/sysctl.conf

add the following then reboot the container
net.ipv6.conf.all.disable_ipv6=1

the container starts up quickly now. hopefully this helps someone in the same situation.

 

[Mike Loiterman]

I’m running into a similar issue - I think?
I have a new Debian 10 container.

The issue I have is that although DHCP assigns both IPv4 and IPv6 addresses, the container only gets an IPv6 DNS for the gateway. In other words looking at /etc/resolv.conf only shows one IPv6 address instead of both IPv4 and IPv6 DNS servers.

About two hours after boot “something” happens and the container must request a DHCP renew which causes the container to lose its IPv4 address.

I can recover by using the host’s console to log in and run “dhclient -4” which then immediately causes the interface to pull IPv4 address, without losing IPv6 address, and pull IPv4 DNS. After this, everything is fine and the VM works fine and subsequent DHCP renews work ok, but issue returns after reboot unless I run the dhclient command.

Confusing and bizarre because I have other Debian 10 containers on the same host that don’t do this...buuuut those did not starts as Debian 10 and were upgraded from 9.

i have the container set to DHCP for both IPv6 and IPv4.

There is a bug here, I think.

 

[nigi]

Hi,
I know this is a very old thread, but the problem still exist on my host.

My config:

net0: name=eth0,bridge=vmbr1,firewall=1,gw=10.40.0.1,hwaddr=26:EA:C0:3A:D4:F8,ip=10.40.0.34/24,type=veth
net1: name=eth1,bridge=vmbr0,firewall=1,hwaddr=1E:5E:08:C3:2D:19,ip=dhcp,type=veth
ostype: debian
unprivileged: 1

I've tried everything mentioned, even disabling IPv6.
But it's always the same, my VM is loosing the IP on net1 within a few minutes or hours.

Is there any solution in the meantime?

 

[adystech]

@nigi did you try adding below to your interface file? in this case eth0 is tied to dhcp. Delete the ipv6 row if you have one.

allow-hotplug eth0