Fault Tolerance with ProCurve 2824's

[aj4525]

Hey Everyone --

Looking for some advice on our current situation. We have a full rack at a data center that contains 1 uplink to the data center distribution layer switch, 2 HP ProCurve 2824's and a handful of servers (I'll leave it at that to keep things simple). Switch A and Switch B are not fault tolerant. If one fails, we'll lose half of our links. We're trying to configure the two HP 2824's so that if we lose one normal operation can resume. We currently have trunk ports configured but this is strictly for increasing bandwidth. For example:

interface 17
name "SERVER-A-BOND0-1"
no lacp
exit
interface 18
name "SERVER-A-BOND0-2"
no lacp
exit

trunk 17-18 Trk4 Trunk

spanning-tree Trk4 priority 4

From what I understand, the switches have to be stacked (Switch A is the commander and Switch B would be a member). After reading the 2824 manual and researching the configuration I honestly feel a bit more confused about how this should be configured correctly. 

I'm assuming once the switches are stacked I can do a similar configuration except interface 17 above would be interface 1 on Switch A and interface 18 would be interface 1 on Switch B. If Switch A or B fails, the server access won't break. The manual specifically mentions using LACP for redundant switches and the "trunk" option for when interoperability is an issue. Would really appreciate any comments as to whether we're going down the right path with the new configuration.

Also, for the switch uplinks we currently have 1 for two switches and the two switches connected to each other. We're going to be getting a second uplink for the other witch. Is there any specific way the distribution layer switch ports would need to be configured (for example LACP)?

[Richard Brodie_1]

Stacking on the traditional Procurve systems is purely for administrative convenience; you cannot treat them as one switch, and split a trunk between them, or anything of that nature.

The only 'true' stacking device with Procurve heritage is the 3800.

[aj4525]

Thank you, I appreciate the reply.  Is that what HP refers to as "Distributed Trunking"? I think in other realms it's referred to as cross-stack EtherChannel.  

[paulgear]

Stacking and distributed trunking are slightly different things, but in terms of LACP they provide equivalent service.  As well as the "true stacking" on the 3800, you can get distributed trunking on the 5400 and 8200 (and possibly 6600/6200?).

IRF on the Comware switches is far preferable, IMO.