Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Executing either command:

sudo service network restart
sudo ifup bond0

Results in the following error (and this is the only error):

Bringing up interface eth0:  RTNETLINK answers: File exists

Here's the config for the involved interfaces:

DEVICE="bond0"
IPADDR="10.100.48.6"
NETMASK="255.255.254.0"
ONBOOT="yes"

DEVICE="eth0"
MASTER="bond0"
SALVE="yes"
HWADDR="78:2B:CB:57:B5:B3"
NM_CONTROLLED="no"
ONBOOT="yes"

DEVICE="eth4"
MASTER="bond0"
SLAVE="yes"
HWADDR="00:1B:21:C9:F4:54"
NM_CONTROLLED="no"
ONBOOT="yes"

The dmesg output suggests eth0 is not ready when bonding driver tries to enslave it:

dmesg | egrep '(eth0)|(bond0)'
bnx2 0000:01:00.0: eth0: Broadcom NetXtreme II BCM5709 1000Base-T (C0) PCI Express found
at mem d4000000, IRQ 36, node addr 78:2b:cb:57:b5:b3
ADDRCONF(NETDEV_UP): bond0: link is not ready
bnx2 0000:01:00.0: eth0: using MSIX
ADDRCONF(NETDEV_UP): eth0: link is not ready
bonding: bond0: Adding slave eth4.
bonding: bond0: enslaving eth4 as a backup interface with a down link.
bnx2 0000:01:00.0: eth0: NIC Copper Link is Up, 100 Mbps full duplex
ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
bond0: link status definitely up for interface eth4, 100 Mbps full duplex.
bonding: bond0: making interface eth4 the new active one.
bonding: bond0: first active interface up!
ADDRCONF(NETDEV_CHANGE): bond0: link becomes ready

If network services are stopped the routing table is empty (command route shows no routes). The eth0 interface can be manually enslaved after bond0 is up it will be chosen as primary and the I can ping devices in the network.

share|improve this question
5  
The RTNETLINK answers: File exists usually indicates that a route already exists and you are trying to add it again. Are you doing anything to clear your existing config before running ifup? With the network in a stopped/down state is your routing table empty? –  Zoredache Feb 24 '12 at 2:10
    
I cannot fully stop the network right now as the machine is remote. When I ifdown bond0 interface the routing table is is empty of the routes relating to to bond0. Eth3 is the network I use to access the machine but that is on another subnet. I've added the output with bond0 down to the main question. –  Lukasz Feb 24 '12 at 5:43
    
Can you possibly setup a VM with an identical network configuration and test then? That is generally what I do when I need to try out networking stuff before I break a remote system. –  Zoredache Feb 25 '12 at 3:12
    
I've got a remote console access to the machine now - I will try your suggestions shortly. –  Lukasz Feb 27 '12 at 16:44
1  
I did a network restart from remote console and while the error persists the bonding driver now enslaves and uses eth0. I did nothing to it. Confusing Linux is Confusing. –  Lukasz Feb 27 '12 at 17:01

1 Answer 1

up vote 1 down vote accepted

Could it simply be a typo? Can you try to turn SALVE into SLAVE for eth0?

share|improve this answer
    
Derp. Thanks I never fixed this. –  Lukasz Jul 5 '12 at 16:17

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.