Re: [suse-sles-e] Problem with network and FW

From: Rafa Grimán (rgriman_at_sgi.com)
Date: Fri Jun 09 2006 - 13:37:27 CEST


From: Rafa Grimán <rgriman@sgi.com>
Date: Fri, 9 Jun 2006 13:37:27 +0200
Message-Id: <200606091337.27287.rgriman@sgi.com>
Subject: Re: [suse-sles-e] Problem with network and FW

Hi :)

El Miércoles, 7 de Junio de 2006 18:22, Stefan Jakobs escribió:
> Hello list,
>
> I'm using SLES9 with bonding on two interfaces. The Server was in
> production and worked fine, till I rebooted yesterday.
> After the reboot the server hangs and all what I see is the output from the
> firewall on the terminal. Something like that:

[..]

> Here is the boot.msg logfile:
>
> #################################################################

[...]

> <notice>start services (network)
> Setting up network interfaces:
> lo
> lo IP address: 127.0.0.1/8
>
> Warning: interface eth1 is not (yet?) active.
> Warning: interface eth0 is not (yet?) active.
> Warning: interface bond0 is not (yet?) active.
> Warning: No interface active (yet?)done eth0 device: Intel
> Corporation 82546EB
> Gigabit Ethernet Controller (Copper) (rev 01)
> eth0 No configuration found for eth0
> unused eth1 device: Intel Corporation 82546EB Gigabit Ethernet
> Controller (Coppe
> r) (rev 01)
> eth1 No configuration found for eth1
> unused
>
> bond0
> bond0 enslaving interfaces: eth0 eth1
> bond0 IP address: 192.168.1.56/24 as bonding master
>
> Warning: interface eth-id-00:07:e9:24:1d:1d is not (yet?)
> active.doneSetting up service
> network . . . . . . . . . . . . . . . .done
> <notice>exit status of (network) is (0)
> #####################################################################
> ## end

States clearly there's no config for eth0 or eth1. Are they correctly setup?

> What I have done before the reboot:
> - Kernel update (this wasn't the fault, I installed the old one already,
> but nothing changed)
> - cron update
> - ocfs2 update
> - cpio update

Doesn't interfere.

> - Adding a rcScript, that runs "ifconfig bond0 add 192.168.11.23" at
> startup

Why do you run this script, it's not needed for channel bonding?

There's a very simple way to setup channel bonding:

1.- YaST -> Network Devices -> Network Device

2.- Configure eth0 and eth1 as DHCP

3.- Configure gateway

4.- Save and exit YaST

5.- cd /etc/sysconfig/network

6.- vi ifcfg-eth0 and replace:
        BOOTPROTO='dhcp'
with
        BOOTPROTO='none'

7.- repeat for eth1

8.- create a file called ifcfg-bond0 with the following:
        BOOTPROTO='static'
        BROADCAST='172.16.0.255'
        IPADDR='172.16.0.1'
        MTU=''
        NETMASK='255.255.255.0'
        NETWORK='172.16.0.0'
        REMOTE_IPADDR=''
        STARTMODE='onboot'
        BONDING_MASTER='yes'
        BONDING_MODULE_OPTS='miimon=100 mode=1 use_carrier=1'
        BONDING_SLAVE0='eth0'
        BONDING_SLAVE1='eth1'

Change BROADCAST, IPADDR, NETMASK, NETWORK and BONDING_MODULE_OPT to match
your needs.

9.- Then add in /etc/modprobe.conf.local

        alias bond0 bonding
        options bonding miimon=100 mode=1 use_carrier=1

The last line must match BONDING_MODULE_OPTS.

10.- rcnetwork restart

That should get you going :)

HTH

   Rafa

-- 
"Even paranoids have enemies."
Rafa Grimán
Systems Engineer
Silicon Graphics Spain
Santa Engracia, 120 - Planta Baja
28003 Madrid
Spain
Tel: +34 91 3984200
Tel: +34 91 3984201
Móvil: +34 628 117 940
http://www.sgi.com
OpenWengo: rgriman
Skype:     rgriman
---------------------------------------------------------------------
To unsubscribe, e-mail: suse-sles-e-unsubscribe@suse.com
For additional commands, e-mail: suse-sles-e-help@suse.com


This archive was generated by hypermail 2.1.7 : Fri Jun 09 2006 - 13:37:48 CEST