Resilient link from a HPE 5130EI IRF switches ad two aruba 2930F Switch

[Francesco_R]

First of all, thanks for your support/suggestions.

I have this situation: 

6 HPE 5130 (JG937A) switch linked with IRF as core switch and, in another floor, n° 2 aruba 2930F (JL262A)

i have 4 fiber cable from 5130 rack to 2930 rack, so in my mind i think to link in this way

   5130A <irf> 5130B <irf> 5130C <irf> 5130D <irf> 5130E <irf> 5130F
                                                  |     \  /      |            
                                                  |      / \      |            
                                          2930F-A   2930F-B

on 5130 i have configure a static link aggregation:

1. interface Ten-GigabitEthernet3/0/51
2. port link-type hybrid
3. port hybrid vlan 5 to 6 9 tagged
4. port hybrid vlan 1 untagged
5. port link-aggregation group 4
6. #
7. interface Ten-GigabitEthernet3/0/52
8. port link-type hybrid
9. port hybrid vlan 5 to 6 9 tagged
10. port hybrid vlan 1 untagged
11. port link-aggregation group 5
12. #
13. interface Ten-GigabitEthernet4/0/51
14. port link-type hybrid
15. port hybrid vlan 5 to 6 9 tagged
16. port hybrid vlan 1 untagged
17. port link-aggregation group 4
18. #
19. interface Ten-GigabitEthernet4/0/52
20. port link-type hybrid
21. port hybrid vlan 5 to 6 9 tagged
22. port hybrid vlan 1 untagged
23. port link-aggregation group 5

on 2930F i configured a static, L3-Based, trunk with port 49 and 50 on every switch.

But when i phisically connect then fiber cable, all network go down ad on console of 5130 i see a lot of message of STP learning states. 

Works when i remove the doble link:

   5130A <irf> 5130B <irf> 5130C <irf> 5130D <irf> 5130E <irf> 5130F
                                                  |     \                    
                                                  |       \                  
                                          2930F-A   2930F-B

No cable connection between 2930F-A and 2930-B

what am i doing wrong?

[parnassus · Accepted solution]

Why static? why Hybrid? Stay away from Hybrid and go simple, go Dynamic (LACP LAG) the topology is OK, not the configuration.

VLAN trunk meaning in Comware (VLANs tagging) is not like on ArubaOS-Switch (port aggregation)...this can cause misconfiguration let you kick in unnecessary (Hybrid) mode on Comware side. ArubaOS-Switch side is somewhat very simple (trunk ethernet 49,50 trk1 lacp for Dynamic LAG = LACP and then tagging required VLANs on trk1, which, by default, is untagged in VLAN 1).

Reset and follow this KB Support Article:
https://support.hpe.com/hpsc/doc/public/display?docId=emr_na-c03503743

There is also a nice LACP presentation on Aruba Airheads...it will shed some lights on LAG on Comware versus ArubaOS-Switch (HP ProVision).

[Francesco_R]

Thanks Parnassus, i'll change the config this weekend and i'll post the result