LACP Bonding not Working

[pmr]

I have a machine here at work which we're trying to get configured with dual ethernet cables. Unfortunately the router on our rack is controlled by a central IT department, and I don't have administrative privileges on it.

I put in a request via our institution's ticketing system, and an IT person claims to have set up LACP bonding on the ports I requested. On the given server I ran modprobe bonding mode=4. My understanding is that 4 indicates LACP. I then configured our network cards as follows:

Code: Select all

[root@localhost network-scripts]$ cat ifcfg-bond0 
NAME="System bond0"
DEVICE=bond0
NM_CONTROLLED=no
ONBOOT=yes
BOOTPROTO=dhcp
USRCTL=no
DEFROUTE=yes
PEERDNS=yes
PEERROUTES=yes

[root@localhost network-scripts]$ cat ifcfg-p1p1
DEVICE=p1p1
BOOTPROTO=none
ONBOOT=yes
MASTER=bond0
SLAVE=yes
USERCTL=no
NM_CONTROLLED=no[code]

[root@localhost network-scripts]$ cat ifcfg-p1p2
DEVICE=p1p2
BOOTPROTO=none
ONBOOT=yes
MASTER=bond0
SLAVE=yes
USERCTL=no
NM_CONTROLLED=no
[/code]

And lastly, here's the output of /proc/net/bonding/bond0

Code: Select all

Ethernet Channel Bonding Driver: v3.6.0 (September 26, 2009)

Bonding Mode: IEEE 802.3ad Dynamic link aggregation
Transmit Hash Policy: layer2 (0)
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0

802.3ad info
LACP rate: slow
Aggregator selection policy (ad_select): stable
Active Aggregator Info:
	Aggregator ID: 5
	Number of ports: 1
	Actor Key: 17
	Partner Key: 1
	Partner Mac Address: 00:22:55:d6:e9:00

Slave Interface: p1p1
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:10:18:5d:ce:c0
Aggregator ID: 5
Slave queue ID: 0

Slave Interface: p1p2
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:10:18:5d:ce:c2
Aggregator ID: 6
Slave queue ID: 0

Something feels wrong here, because p1p1 and p1p2 get different "Aggregator ID" numbers, and overall the bond says "Number of ports: 1", which I think should be 2. Indeed, when I try slam the machine with network traffic from multiple different locations at once it never goes past 1Gb/s, even though with the boding it should support 2Gb/s (across multiple parallel TCP streams).

I'm not sure what I'm doing wrong here. Is there something wrong on my end, or is there something misconfigured on the switch?

[TrevorH]

This does now look more healthy to me than it did when you were posting on IRC - "Transmit Hash Policy: layer2 (0)" looks better for example.

Personally, I would add the BONDING_OPTS="mode=4 miimon=100" to your ifcfg-bond0 and modprobe -r bonding with the link down then bring it back up and if it still fails then it's time to go back to network admin and ask them to take another look

Oh, one other thing that might be worth trying, if NetworkManager is installed then I would remove it completely, even with NM_CONTROLLED=no I still don't trust it. I find that yum remove NetworkManager\* tends to stop it from interfering!