LACP redundant links HP procruve 5412zl to Comware A5500 stack

[bigredcherokee]

I have a 5412zl with two J9536a cards that have the 10gig mini gbic cables running to a stack of A5500 switches. Right now I have only have one link up between the two, but I want to add another for redundancy.

On the 5412zl I have it configured as such. Working with ports B22 and H22

trunk B22,H22 trk5 lacp

On all my vlans I have tagged B22, H22 Trk5

Spanning-tree Trk5 priority 4

On the A5500 I'm working with ports 1/0/30 and 2/0/30

Interface Ten-GigabitEthernet2/0/30
port link-mode bridge
port link-type hybrid
port hybrid vlan 13 23 33 53 73 84 217 tagged
port hybrid vlan 1 untagged

I think I need to add the following.

interface Bridge-Aggregation23

port access vlan all

link-aggregation mode dynamic

Then change 1/0/30 and 2/0/30 to

Interface Gigabitehternet1/0/30 and 2/0/30

port access vlan all

port link-aggeration group 23

[bigredcherokee]

got the documentation I needed.

http://h20564.www2.hpe.com/hpsc/doc/public/display?docId=emr_na-c03503743&sp4ts.oid=4177519

[parnassus · Accepted solution]

Recently HPE published an How-To about how to configure LACP aggregation between a Comware (A and E Switch Series) and a ProVision (now Aruba-OS Switch) switch.

That How-To should be basically considered valid even if your Port Trunking, originating from the HPE 5412 zl, instead of terminating (co-terminating) on a single physical Switch (as in the How-To given example) is going to terminate on the IRF Stack (which is a "virtual" Logical Switch) with each of its trunk's member link terminating on a different IRF Member (Member 1 Port 1/0/30 and Member 2 Port 2/0/30)...this shouldn't be an issue di per sè (I mean: I don't think Distributed Trunking - known as dt-lacp versus the standard lacp - is required in this case since the IRF Stack should be seen by the HPE 5412 zl really as a single physical Switch).

I think your procedure (ProVision side looks good to me, apart that reference to Spanning Tree priority on the LAG) it's substantially right but, as is written on that How-To, on Comware switches the order of execution of system commands is more important, so I think you should execute something like (without changing any single LAG member ports setting after your LAG is finally configured on Step 3):

When configuring a link aggregation on A-Series switches (Comware)
the order of execution is very important.

The following are the configuration steps.

Step 1: Create the Link Aggregation Interface
Step 2: Associate the Member Ports to the Link Aggregation Interface
Step 3: Make any additional changes to the Link Aggregation Interface

Step 1: Create the Link Aggregation and set the mode to Dynamic (LACP)

[irfstack] interface Bridge-Aggregation 23
[irfstack-Bridge-Aggregation23] link-aggregation mode dynamic

Step 2: Associate the the Member Ports to the Link Aggregation Interface

[irfstack] interface Ten-GigabitEthernet1/0/30
[irfstack-Ten-GigabitEthernet1/0/30] port link-aggregation group 23

[irfstack] interface Ten-GigabitEthernet2/0/30
[irfstack-Ten-GigabitEthernet2/0/30] port link-aggregation group 23
 
Step 3: set the Link Aggregation Interface with Port Link Type to type Trunk
and set which VLANs to permit (all or just a list)

[irfstack] interface Bridge-Aggregation 23
[irfstack-Bridge-Aggregation23] port link-type trunk
[irfstack-Bridge-Aggregation23] port trunk permit vlan 1 13 23 33 53 73 84 217

[parnassus]

Argh...only 22 minutes after you!