netdev - Deleting an alias causes rest to get deleted

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Date:	Mon, 02 Jul 2012 22:54:28 +0300
From:	Volkan Yazıcı <volkan.yazici@...il.com>
To:	netdev@...r.kernel.org
Subject: Deleting an alias causes rest to get deleted

Hi!

I observe an IP aliasing anomaly that occurs when I try to delete an IP 
alias from an interface. That is, when I delete the first address in a 
set of IP aliased addresses assigned according to a particular subnet, 
rest of the aliases get deleted as well. Check out the below snippet.

    $ *for I in `seq 1 6`; do sudo ip addr add 192.168.2.$I/29 dev eth0;
    done*
    $ ip addr list
    1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
         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
    2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast
    state UP qlen 1000
         link/ether 00:24:54:b9:1c:f8 brd ff:ff:ff:ff:ff:ff
    *inet 192.168.1.200/24 brd 192.168.1.255 scope global eth0**
         inet 192.168.2.1/29 scope global eth0
         inet 192.168.2.2/29 scope global secondary eth0
         inet 192.168.2.3/29 scope global secondary eth0
         inet 192.168.2.4/29 scope global secondary eth0
         inet 192.168.2.5/29 scope global secondary eth0
         inet 192.168.2.6/29 scope global secondary eth0*
         inet6 fe80::224:54ff:feb9:1cf8/64 scope link
            valid_lft forever preferred_lft forever
    3: wlan0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN qlen 1000
         link/ether e8:39:df:6a:21:2a brd ff:ff:ff:ff:ff:ff
    $ *sudo ip addr del 192.168.2.1/29 dev eth0*
    $ ip addr list
    1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
         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
    2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast
    state UP qlen 1000
         link/ether 00:24:54:b9:1c:f8 brd ff:ff:ff:ff:ff:ff
    *inet 192.168.1.200/24 brd 192.168.1.255 scope global eth0*
         inet6 fe80::224:54ff:feb9:1cf8/64 scope link
            valid_lft forever preferred_lft forever
    3: wlan0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN qlen 1000
         link/ether e8:39:df:6a:21:2a brd ff:ff:ff:ff:ff:ff

Per see, deleting 192.168.2.1/29 causes the rest of the aliased 
interfaces get deleted as well. This is something that is slightly 
documented in the ifconfig manual: /for every scope (i.e. same net with 
address/netmask combination) all aliases are deleted, if you delete the 
first (primary)/. So what is the right way of just deleting the first 
(primary) alias without affecting the rest? If this is a scoping issue, 
is it possible to assign each alias as primary within its own dedicated 
scope?

As a side note, when I first asked this question to Stephen Hemminger 
(he forwarded me to this mailing list) he also told me that "/In Linux 
the interface aliases are really a legacy from the BSD style addressing, 
and don't act the same. It is not common practice to use them./" Is that 
really the case? Because, as you know, IP aliasing is the heart of a 
majority of the high-availability and clustering solutions in Linux. Is 
IP aliasing a really deprecated technology in Linux? Should we avoid 
using it? If so, what do you recommend as an alternative?

Best.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html