Package: ifupdown Version: 0.7.48.1 Severity: normal Tags: upstream, ipv6 Hi,
My system has this:
doublan-1# cat /etc/network/interfaces
auto lo
iface lo inet loopback
auto eth0
iface eth0 inet static
address 137.56.165.54
netmask 255.255.255.0
gateway 137.56.165.1
iface eth0 inet6 static
address 2001:610:1410:480:f35e:2a02::54
netmask 64
iface eth0:0 inet static
address 137.56.165.59
netmask 255.255.255.0
iface eth0:0 inet6 static
address 2001:610:1410:480:f35e:2a02::59
netmask 64
doublan-1# ip a l eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state
UNKNOWN qlen 1000
link/ether 00:50:56:13:01:aa brd ff:ff:ff:ff:ff:ff
inet 137.56.165.54/24 brd 137.56.165.255 scope global eth0
inet6 2001:610:1410:480:f35e:2a02:0:54/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80::250:56ff:fe13:1aa/64 scope link
valid_lft forever preferred_lft forever
Now, if I run
doublan-1# ifup eth0:0
I am happy to get:
doublan-1# ip a l eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state
UNKNOWN qlen 1000
link/ether 00:50:56:13:01:aa brd ff:ff:ff:ff:ff:ff
inet 137.56.165.54/24 brd 137.56.165.255 scope global eth0
inet 137.56.165.59/24 brd 137.56.165.255 scope global secondary eth0:0
inet6 2001:610:1410:480:f35e:2a02:0:59/64 scope global deprecated
valid_lft forever preferred_lft forever
inet6 2001:610:1410:480:f35e:2a02:0:54/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80::250:56ff:fe13:1aa/64 scope link
valid_lft forever preferred_lft forever
However, if I now run:
doublan-1# ifdown eth0:0
I am _not_ happy to get:
doublan-1# ip a l eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state
UNKNOWN qlen 1000
link/ether 00:50:56:13:01:aa brd ff:ff:ff:ff:ff:ff
inet 137.56.165.54/24 brd 137.56.165.255 scope global eth0
inet6 fe80::250:56ff:fe13:1aa/64 scope link
valid_lft forever preferred_lft forever
Only the legacy IPv4 adress survived this action, I am missing:
inet6 2001:610:1410:480:f35e:2a02:0:54/64 scope global
valid_lft forever preferred_lft forever
.
This seems to be caused by differences in IPv4 vs IPv6 handling. In the
IPv4-case, ifdown calls:
ip -4 addr flush dev eth0:0 label eth0:0
which is fine, since only the eth0:0 adresses cary this label.
In the IPv6-case, ifdown calls:
ip -6 addr flush dev eth0:0 scope global
which breaks, since the kernel interprets “eth0:0” as “eth0”.
This bug has been seen with ifupdown version 0.7.8. I cannot easily test it on
a system running 0.7.48.1 (which is in testing/jessie and sid now). However,
since the relevant code hasn't been changed, I'm convinced ifupdown 0.7.48.1 is
broken in the same way.
Thanks, Bye,
Joost
--
Joost van Baal-Ilić http://abramowitz.uvt.nl/
Tilburg University
mailto:joostvb.uvt.nl The Netherlands
signature.asc
Description: Digital signature
