Hello,

I am seeing an issue on a few servers where it doesn't appear all NICs in the 802.3ad lag are all operating at the same level. A few of the servers have two bonds each with two NICs in each bond. I have two NFS servers that each has 1 bond with 3 NICs. All are RHEL5 x64 2.6.18.

Here's one of my bonds:
[root@server ~]# cat /proc/net/bonding/bond0
Ethernet Channel Bonding Driver: v3.4.0-1 (October 7, 2008)

Bonding Mode: IEEE 802.3ad Dynamic link aggregation
Transmit Hash Policy: layer2 (0)
MII Status: up
MII Polling Interval (ms): 80
Up Delay (ms): 0
Down Delay (ms): 0

802.3ad info
LACP rate: slow
Active Aggregator Info:
Aggregator ID: 3
Number of ports: 3
Actor Key: 17
Partner Key: 10010
Partner Mac Address: 00:1b:ed:80:17:c0

Slave Interface: eth0
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 2
Permanent HW addr: 00:1d:09:71:97:3f
Aggregator ID: 3

Slave Interface: eth1
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 3
Permanent HW addr: 00:1d:09:71:97:41
Aggregator ID: 3

Slave Interface: eth3
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 2
Permanent HW addr: 00:15:17:6c:e4:89
Aggregator ID: 3


Here's a snippet of sar:
[root@server ~]# sar -n DEV 1 5

Average: IFACE rxpck/s txpck/s rxbyt/s txbyt/s rxcmp/s txcmp/s rxmcst/s
Average: lo 0.40 0.40 33.67 33.67 0.00 0.00 0.00
Average: eth0 1924.25 0.00 1574977.35 0.00 0.00 0.00 0.00
Average: eth1 7034.87 524.25 41138060.92 3758143.29 0.00 0.00 0.00
Average: eth2 4608.82 3578.56 11813732.26 17426152.71 0.00 0.00 0.00
Average: eth3 352.91 7568.94 174972.34 31082267.33 0.00 0.00 0.00
Average: eth4 4613.83 3667.94 12243683.77 17352368.34 0.00 0.00 0.00
Average: eth5 4533.87 3526.25 11913877.35 17389009.22 0.00 0.00 0.00
Average: sit0 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Average: bond0 9312.02 8093.19 42888010.62 34840410.62 0.00 0.00 0.00

I couldn't read that without having a stroke so here it is broken down in MB/s by interface:
rx MB/s tx MB/s
eth0 1.5 0.0
eth1 39.2 3.6
eth3 0.2 29.6
bond0 40.9 33.2

I think the reason why I see one interface dominating RX and another dominating TX is due to the xmit_hash_policy but there are three hosts that use this particular server for network traffic. That's 3 different physical mac addresses. The layer2 algorithm should be fine in that situation I would think. What am I missing? Would I just be better off with balance-rr?

Thanks!

You could try balance-alb (mode 6), this will enable receive load balancing as well. Try reading bonding.txt and see what suits your traffic patterns

hth

This is expected. Need an even number of interfaces. I saw a reasonable even distribution with two interfaces.

It appears mode 6 does things more evenly across 3 interfaces. If I ever go to 4, I will try 802.3ad again.