Article ID: 112066, created on Aug 25, 2011, last review on May 11, 2014

  • Applies to:
  • Virtuozzo hypervisor 5.0

Symptoms

A container based on the centos-5-x86_64 OS template may have issues with restarting the dhcp6c daemon.

If the container has a bridged interface with a DHCPv6-assigned IPv6 address, the address disappears after adding the IPv6 address to the host-routed interface venet0:
~# vzctl exec 111 ip a l 
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue 
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 
    inet 127.0.0.1/8 scope host lo 
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever 
3: venet0: <BROADCAST,POINTOPOINT,NOARP> mtu 1500 qdisc noop 
    link/void 
5: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue 
    link/ether 00:18:51:b4:a8:fb brd ff:ff:ff:ff:ff:ff 
    inet 192.168.3.7/16 brd 192.168.255.255 scope global eth0 
    inet6 fd00::75e1:8eb2:bd95:4789/64 scope global 
       valid_lft forever preferred_lft forever 
    inet6 fe80::218:51ff:feb4:a8fb/64 scope link 
       valid_lft forever preferred_lft forever 

~# vzctl set 111 --ipadd fd00::4:17 --save 
Adding IP address(es): fd00::4:17 
Shutting down interface eth0: [ OK ] 
Shutting down loopback interface: [ OK ] 
SIOCGIFFLAGS: No such device 
Bringing up loopback interface: [ OK ] 
Bringing up interface eth0: 
Determining IP information for eth0... done. 
[ OK ] 
Bringing up interface venet0: Global IPv6 forwarding is disabled in configuration, but not currently disabled in kernel 
Please restart network with '/sbin/service network restart' 
[ OK ] 
Saved parameters for Container 111
 
~# vzctl exec 111 ip a l 
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue 
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 
    inet 127.0.0.1/8 scope host lo 
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever 
3: venet0: <BROADCAST,POINTOPOINT,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue 
    link/void 
    inet 127.0.0.1/32 scope host venet0 
    inet6 fd00::4:17/128 scope global 
       valid_lft forever preferred_lft forever 
5: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue 
    link/ether 00:18:51:b4:a8:fb brd ff:ff:ff:ff:ff:ff 
    inet 192.168.3.7/16 brd 192.168.255.255 scope global eth0 
    inet6 fe80::218:51ff:feb4:a8fb/64 scope link 
       valid_lft forever preferred_lft forever

Cause

The root cause of the issue is a bug in the stock initscripts package:
Red Hat bug #683952 dhcp6c client daemon doesn't stop if DHCPV6C_OPTIONS undefined

Resolution

The workaround is to restart the container.

c662da62f00df94fd77ba7a2c9eff4b4 a26b38f94253cdfbf1028d72cf3a498b 2897d76d56d2010f4e3a28f864d69223

Email subscription for changes to this article
Save as PDF