Optimal configuration on XS728T for RN626X with LACP bonding

[Tonkinite]

My 10G ports on my RN626X are LACP bonded and plugged in to my XS728T and this appears to work fine without any special configuration on the switch, but am concerned that I am getting suboptimal throughput.  How should I be configuring the ports on the XS728T to optimize throughput in this situation? The ReadyNAS ports are configured with a MTU of 9000 (aware that 10G can go larger, but 9000 is compatible with the other devices on the Switch, minimal reassembly, etc.).

Advice?

[DaneA · Accepted solution]

@Tonkinite,

Kindly try the steps below:

1. On the RN626X, set the Teaming Mode and Hash Type to 802.11ad and Layer 2 respectively.

2. On the XS728T, set the LAG Type to LACP.

Let us know the results.

Regards,

DaneA

NETGEAR Community Team

[DaneA]

Hi Tonkinite,

Kindly answer the questions below:

a. What is the Teaming Mode and Hash Type configured on the RN626X? 

b. What is the LAG Type configured on the XS728T?

c. What tests did you perform that you came up that you're getting suboptimal throughput with the LACP configured between the RN626X and the XS728T?

d. What is the current firmware of the XS728T?

e. What is the current firmware version of the RN626X? 

Regards,

DaneA

NETGEAR Community Team

[Tonkinite]

Answers:

1. 802.11ad, L2+L3
2. None at present, have tried LAG and LACP settings (but just the standard setups), and the link stopped working entirely.
3. None, I'm asking the question because I don't think I have the proper equipment to test.
4. Version 6.5.1.26
5. Version 6.6.1

[DaneA · Accepted solution]

@Tonkinite,

Kindly try the steps below:

1. On the RN626X, set the Teaming Mode and Hash Type to 802.11ad and Layer 2 respectively.

2. On the XS728T, set the LAG Type to LACP.

Let us know the results.

Regards,

DaneA

NETGEAR Community Team

[Tonkinite]

That worked beautifully, thanks!

[Tonkinite]

So, the XS728T can't handle L2+L3 hashing or other variants beyond the standard L2, that seems shortsighted, to say the least.

[LaurentMa]

Hi

Web Smart Managed Switches only support Layer 2 hashing for LAGs. That is efficient, only issue is when all packets from servers have same destination MAC address = the switch routing interface MAC, when routing enabled on the switch and when clients and servers are in different subnets. If your servers and clients are in same subnet, then L2 (MAC) destination hashing algorithm is good in most cases.

Fully Managed Switches, for instance M4300 series, offer selectable, granular hashing algorithms with L2 (MAC) , L3 (IP) and L4 (TCP/UDP) source and destination address parsing. This offers maximum interoperability.

The XS728T L2 hashing is indicated in datasheet, which is the repository for supported features (and limitations).

Hope this helps.