CentOS 4.4 network bonding issues on PowerEdge 860

Issues related to configuring your network

CentOS 4.4 network bonding issues on PowerEdge 860

Postby malign » 2007/01/23 19:32:23

Hello,

I have done a quick search through this forum for other bonding issues but haven't found anything relevant to my problem(s) so I figured I'd start a new thread and see where it took me.

I have a brand new PowerEdge 860 box just in. It has a dual-port onboard Broadcom gigabit NIC and an add-in Intel dual-port gigabit NIC. In the interest of maximum throughput my boss wanted me to set it up using bandwidth aggregation to turn the 4 interfaces into two. We had done some research using Intel ANS a few years back but that seems to be obsolete and unsupported now. I've seen a few websites detailing how to set up bonding and I've managed to get the two Broadcom interfaces to bond in a round-robin fashion under "bond0", but beyond that I cannot do the following:

1) I cannot get the second bonding interface to come up. I can take eth2 and eth3 and enslave them to a "bond1" interface, but when I do a "service network restart", it says "bonding device bond1 does not seem to be present, delaying initialization". When I set up eth2 and eth3 and do an "ifup" on each, it says it is enslaving the device to bond1, but trying to bring up bond1 fails.

2) I cannot seem to pass in any bonding options in the modprobe.conf file like some guides I've read have said to add an "options" line into modprobe.conf specifying the mode and the "miimon" for failover monitoring threshold. I add these in but looking at the /proc/net/bonding/bond0 file shows that the options are not being loaded as it is going with its default values. I've put the "options" directive all over the place in the file, as well as the "alias bond# bonding" directives as well. Before the eth# declarations AND after.

I am attaching a large volume of output from the server below


"service network restart" output:
[font=Courier]
Shutting down interface bond0: [ OK ]
Shutting down loopback interface: [ OK ]
Setting network parameters: [ OK ]
Bringing up loopback interface: [ OK ]
Bringing up interface bond0: [ OK ]
Bringing up interface bond1: bonding device bond1 does not seem to be present, delaying initialization.
[FAILED]
[/font]


The following is a tail of /var/log/messages:
[font=Courier]
Jan 23 13:10:47 xxx kernel: bonding: bond0: enslaving eth0 as an active interface with an up link.
Jan 23 13:10:47 xxx ifup: Enslaving eth1 to bond0
Jan 23 13:10:47 xxx kernel: bonding: bond0: enslaving eth1 as an active interface with an up link.
Jan 23 13:10:47 xxx network: Bringing up interface bond0: succeeded
Jan 23 13:10:47 xxx ifup: bonding device bond1 does not seem to be present, delaying initialization.
Jan 23 13:10:47 xxx network: Bringing up interface bond1: failed
Jan 23 13:10:49 xxx kernel: tg3: eth0: Link is up at 100 Mbps, full duplex.
Jan 23 13:10:49 xxx kernel: tg3: eth0: Flow control is on for TX and on for RX.
Jan 23 13:10:49 xxx kernel: tg3: eth1: Link is up at 100 Mbps, full duplex.
Jan 23 13:10:49 xxx kernel: tg3: eth1: Flow control is on for TX and on for RX.
[/font]


The following are the ifcfg-eth0 through 3 config files:
[font=Courier]
DEVICE=eth0
ONBOOT=yes
BOOTPROTO=none
MASTER=bond0
SLAVE=yes
USERCTL=yes

DEVICE=eth1
BOOTPROTO=static
MASTER=bond0
SLAVE=yes
USERCTL=yes
TYPE=Ethernet
ONBOOT=yes

DEVICE=eth2
BOOTPROTO=none
MASTER=bond1
SLAVE=yes
USERCTL=no
TYPE=Ethernet
ONBOOT=yes

DEVICE=eth3
BOOTPROTO=none
MASTER=bond1
SLAVE=yes
USERCTL=no
TYPE=Ethernet
ONBOOT=yes
[/font]


The following are the ifcfg-bond0 and bond1 files, and modprobe.conf:
[font=Courier]
DEVICE=bond0
BOOTPROTO=none
ONBOOT=yes
NETWORK=xxx.yyy.zzz.0
NETMASK=255.255.255.0
IPADDR=xxx.yyy.zzz.183
USERCTL=no

DEVICE=bond1
BOOTPROTO=none
ONBOOT=yes
NETWORK=xxx.yyy.zzz.0
NETMASK=255.255.255.0
IPADDR=xxx.yyy.zzz.184
USERCTL=no

alias bond0 bonding
options bond0 mode=6 miimon=100
alias bond1 bonding
options bond1 mode=6 miimon=100
alias eth0 tg3
alias eth1 tg3
alias eth2 e1000
alias eth3 e1000
alias scsi_hostadapter mptbase
alias scsi_hostadapter1 mptscsi
alias scsi_hostadapter2 mptfc
alias scsi_hostadapter3 mptspi
alias scsi_hostadapter4 mptsas
alias scsi_hostadapter5 mptscsih
alias usb-controller ehci-hcd
alias usb-controller1 uhci-hcd
[/font]


And finally, the output of /proc/net/bonding/bond0:
[font=Courier]
Ethernet Channel Bonding Driver: v2.6.3 (June 8, 2005)

Bonding Mode: load balancing (round-robin)
MII Status: up
MII Polling Interval (ms): 0
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth0
MII Status: up
Link Failure Count: 0
Permanent HW addr: abcxyz

Slave Interface: eth1
MII Status: up
Link Failure Count: 0
Permanent HW addr: abcxyz
[/font]
malign
 
Posts: 4
Joined: 2005/07/08 13:41:51

Re: CentOS 4.4 network bonding issues on PowerEdge 860

Postby malign » 2007/01/23 19:40:38

Just as a follow-up, attached is my "lsmod" output as well:

[font=Courier]
Module Size Used by
md5 8129 1
ipv6 242657 12
ipt_REJECT 10689 1
ipt_state 5953 2
ip_conntrack 46085 1 ipt_state
iptable_filter 6977 1
ip_tables 22721 3 ipt_REJECT,ipt_state,iptable_filter
dm_mirror 31901 0
dm_mod 60741 1 dm_mirror
button 10705 0
battery 12997 0
ac 8901 0
uhci_hcd 32857 0
ehci_hcd 32325 0
e1000 109369 0
tg3 101061 0
bonding 65128 0
ext3 118857 1
jbd 59609 1 ext3
mptscsih 5569 0
mptsas 13389 3 mptscsih
mptspi 13261 1 mptscsih
mptfc 12489 1 mptscsih
mptscsi 43997 3 mptsas,mptspi,mptfc
mptbase 61345 4 mptsas,mptspi,mptfc,mptscsi
sd_mod 20545 3
scsi_mod 117709 5 mptsas,mptspi,mptfc,mptscsi,sd_mod
[/font]
malign
 
Posts: 4
Joined: 2005/07/08 13:41:51

Re: CentOS 4.4 network bonding issues on PowerEdge 860

Postby TyRoden » 2007/02/07 14:21:42

I am having this identical problem on an HP server as well.
One dual NIC card is a e1000 driver and the other is a tg3 driver.

I have hunted all over the web trying to find a solution to this one, and it seems that sometimes people do post back a message saying that they had found the solution, yet it never seems to do it in my case when I paste in the 'fix'. Even attempted the manual installation process to no avail.

Of the eth0-eth3 devices, it does not matter which pair I attempt to make bond1 - whatever configuration chosen, the cards setup to be in bond1 are the ones that fail. Either pair can be made into a successful bond0 - it's just when adding bond1, so i cannot even blame the different network drivers ;-)

Any ideas out there?
TyRoden
 
Posts: 1
Joined: 2007/02/07 14:03:56

Re: CentOS 4.4 network bonding issues on PowerEdge 860

Postby davidn » 2007/02/15 14:21:32

I have the same problem on a Dell PE 2650. It looks like the bonding1 option does not get loaded but am not sure why. When run lsmod on a RH ES3 box (same Dell PE 2650) it shows both bonding and bonding1 loaded. When I try to manual load bonding1 on CentOS 4 it fails.
davidn
 
Posts: 2
Joined: 2007/02/15 14:16:30

Re: CentOS 4.4 network bonding issues on PowerEdge 860

Postby davidn » 2007/02/15 16:08:23

I have been able to get this working with the below config but had to use rmmod to delete both network card drivers (e100 and tg3) and the bonding driver from the kernel before rebooting / restarting networking daemon.

# cat /etc/modprobe.conf
alias eth0 e1000
alias eth1 e1000
alias eth2 tg3
alias eth3 tg3

alias bond0 bonding
options bond0 miimon=100 mode=1 max_bonds=2 primary=eth0
alias bond1 bonding
options bond1 miimon=100 mode=1 primary=eth1

# cat ifcfg-bond0
DEVICE=bond0
BOOTPROTO=none
ONBOOT=yes
NETWORK=x.x.x.x
NETMASK=255.255.255.0
IPADDR=x.x.x.x
USERCTL=no

# cat ifcfg-bond1
DEVICE=bond1
BOOTPROTO=none
ONBOOT=yes
NETWORK=x.x.x.x
NETMASK=255.255.255.0
IPADDR=x.x.x.x
USERCTL=no

# cat ifcfg-eth0
DEVICE=eth0
BOOTPROTO=none
ONBOOT=yes
MASTER=bond0
SLAVE=yes
USERCTL=no

# cat ifcfg-eth1
DEVICE=eth1
BOOTPROTO=none
ONBOOT=yes
MASTER=bond1
SLAVE=yes
USERCTL=no

# cat ifcfg-eth2
DEVICE=eth2
BOOTPROTO=none
ONBOOT=yes
MASTER=bond0
SLAVE=yes
USERCTL=no

# cat ifcfg-eth3
DEVICE=eth3
BOOTPROTO=none
ONBOOT=yes
MASTER=bond1
SLAVE=yes
USERCTL=no

Hope it helps
- David
davidn
 
Posts: 2
Joined: 2007/02/15 14:16:30

Re: CentOS 4.4 network bonding issues on PowerEdge 860

Postby zioalex » 2007/08/22 12:38:41

ThanX Davidn,
I've solved the same problem on CentOS5 ...

But Have you any idea of why?

ZioAlex
zioalex
 
Posts: 3
Joined: 2007/08/22 12:35:07

Re: CentOS 4.4 network bonding issues on PowerEdge 860

Postby celsowebber » 2008/04/02 18:53:09

Hello,

I'm finding this myself with problems similar than yours, but with CentOS 5 (also with RHEL 5).

After digging a little bit, I found that it worked when I configured /etc/modprobe.conf like this:

alias bond0 bonding
install bond0 /sbin/modprobe bonding -o bonding0 mode=1 miimon=100

If you need ore than one bonding interface, just repeat the lines above for the other bonding interfaces.

I found this solution on this blog: http://bisscuitt.blogspot.com/2007/10/r ... nding.html

According to the author, this was the only consistent way of making this work.

Regards,

Celso.
celsowebber
 
Posts: 2
Joined: 2006/02/15 23:22:50


Return to CentOS 4 - Networking Support

Who is online

Users browsing this forum: No registered users and 1 guest