W2008 R2 Network config with P4300 - sw2910al

[Fred Blum]

I am looking for a network best practices setup with W2008 R2 server (future Hyper-V role) with a P4300 san (two nodes) and two ProCurve 2910al-24G switches.

We have 6 nic's of which two are teamed in the DL380 G7 and cross connected to the two switches and NICS on the two nodes in an isolated subnet. Node NICS are bonded. Connectivity is there, Volumes created, formatted and drive letter accessible on the server.

I want to understand the best practise setup, which should be setting up load balancing/fault tolerance, jumbo package size and flow control on the NIC's and the switch.

There are several team types on HP's Network Config Utility but adaptive load balancing with fault tolerance as I have set on the node is not present. Which team type to choice from corresponding to that of the SAN? Tried 802ad dynamic load balancing and fault tolerance but I then lose connectity to the SAN. Jumbo package sizes are still default.

Do I need to configure the 2910al switches also? Connect the switches to eachother also? How do the ports on the switches need to be configured?

Is there a Best practices walkthrough for this P4300 SAN - ProCurve 2910al-24G setup on W2008 R2 (future Hyper-V)?

[Fred Blum]

My HP NCU no longer mentions ALB for this reason:
"Transmit Load Balancing with Fault Tolerance (TLB)
Transmit Load Balancing mode, previously known as Adaptive Load Balancing (ALB), incorporates all the features
of NFT, plus Transmit Load Balancing. In this mode, two to eight ports may be teamed together to function as a
single virtual network port. The load-balancing algorithm used in TLB allows the server to load balance traffic
transmitted from the server. However, traffic received by the server is not load balanced, meaning the Primary
teamed port is responsible for receiving all traffic destined for the server (refer to Figure 4-12). As with NFT, there
are two types of team members, Primary and Non-Primary ports. The Primary port transmits and receives frames
and the non-Primary ports only transmit frames."
http://h10032.www1.hp.com/ctg/Manual/c01750150.pdf

So I need to set it to TLB to correspond to the node bond config with no need to config the switches. Is there a need to connect the switches to eachother?

Isn't link aggregation with transmit/receive balancing a beter practise? Walkthrough avalaible?

Do the nic's have to have the iSCSI device setting enabled? Currently not enabled. Is that TCP offloading? How does it change the W2008 R2 configuration?

[Fred Blum]

I am running an IOMeter test against the server's local SAS 15K HD and comparing to the SAN. I am seeing 100% reads 140MBPS local HD versus 67MBPS SAN Raid 5.

This is with TLB, defailt package size and no additional switch settings. Dissapointing. I've read that 120MB will be the max. Is that with 802.ad link aggegate, jumbo, and flow control?