Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[xianthax]

I have a 1216T rev D2 and would like to know the best way to use its trunking feature for bonded links in linux using the bonding module.

The switch doesn't seem to support 802.3ad, or at least enabling bonding in this mode generates errors.

In this case i'm wondering what is the best bonding mode to use.  The goal is to bond the 2 gigabit links on my workstation and the 2 gigabit links on my file server to improve transfer speeds in both directions between these systems.  Before you ask both devices are more than capable of saturating the full the 2 gigabit bandwidth that would be created, even on sustained reads/writes.

I do need this bandwidth for a single stream, for example working on a large file over NFS.  As such bonding modes that only spread connections to various slave links won't be helpful.

As i don't know how the switching trunking feature works, i'm not sure which is the best option to interact with it.

The available bonding modes that include load balancing are:

mode=0   (balance-rr) 
Round-robin policy: Transmit packets in sequential order from the first available slave through the last. This mode provides load balancing and fault tolerance.

mode=2   (balance-xor)
XOR policy: Transmit based on [(source MAC address XOR'd with destination MAC address) modulo slave count]. This selects the same slave for each destination MAC address. This mode provides load balancing and fault tolerance.

mode=4   (802.3ad)
IEEE 802.3ad Dynamic link aggregation. Creates aggregation groups that share the same speed and duplex settings. Utilizes all slaves in the active aggregator according to the 802.3ad specification.

mode=5   (balance-tlb)
Adaptive transmit load balancing: channel bonding that does not require any special switch support. The outgoing traffic is distributed according to the current load (computed relative to the speed) on each slave. Incoming traffic is received by the current slave. If the receiving slave fails, another slave takes over the MAC address of the failed receiving slave.

mode=6 (balance-alb)
Adaptive load balancing: includes balance-tlb plus receive load balancing (rlb) for IPV4 traffic, and does not require any special switch support. The receive load balancing is achieved by ARP negotiation. The bonding driver intercepts the ARP Replies sent by the local system on their way out and overwrites the source hardware address with the unique hardware address of one of the slaves in the bond such that different peers use different hardware addresses for the server.

It would appear that mode 6 may be effective without switch support, however you do pay the overhead of additional ARP negotiation.  mode 0 would appear to be for TX only, but this mode sets both NIC's to the same MAC address, perhaps this combined with the trunking feature in the switch would balance reception as well?

Thanks for any help.

[Fatman]

As this switch doesn't support LACP, you are not going to get LACP bonding working in Linux.  If you hook up a witch that supports LACP there is some hope.  The static link aggregation this switch supports is only guaranteed to work between switches of the same model.