VSF mit HP Aruba 5406 and Trunk

[AndyZ1]

We have a network with 2 Core HP switches configured as VSF.
Each switch has an identical design.

All access switches are redundantly connected
(e.g. Core1 - Port 1/A1 and Core2 - Port 2/A1)

LACP and Trunk is configured for 1/C11 and 2/C11
(trunk 1/C11,2/C11 trk26 lacp)

Currently 4 VLANs are defined:

vlan 1
   name "DEFAULT_VLAN"
   no untagged 1/A1,2/A1
   untagged
 1/A2-1/A12,1/C12-1/C24,1/D1-1/D24,1/E7,1/F1-1/F7,2/A2-2/A12,2/C12-2/C24,2/E7-2/
E8,2/F1-2/F7,Trk101-Trk124,Trk130-Trk135
   tagged Trk10-Trk26
   no ip address
   exit
vlan 2002
   name "MGMT"
   untagged 1/A1,2/A1
   tagged Trk10-Trk26
   ip address 10.20.2.254 255.255.254.0
   exit
vlan 2020
   name "VOIP"
   tagged Trk10-Trk26
   ip address 10.20.20.254 255.255.254.0
   exit
vlan 2033
   name "WIFI extern"
   tagged Trk10-Trk26
   ip address 10.20.33.254 255.255.255.0
   exit

more infos (show tech all)
spanning-tree Trk133 priority 4 bpdu-filter
  Port   | Name                             Type       | Group  Type    
  ------ + -------------------------------- ---------- + ------ --------
  1/A18  |                                  100/1000T  | Trk133 LACP    
  1/A19  |                                  100/1000T  | Trk133 LACP    
  1/A20  |                                  100/1000T  | Trk133 LACP
  2/A18  |                                  100/1000T  | Trk133 LACP    
  2/A19  |                                  100/1000T  | Trk133 LACP    
  2/A20  |                                  100/1000T  | Trk133 LACP

                          | Intrusion                           MDI  Flow Bcast
  Port         Type       | Alert     Enabled Status Mode       Mode Ctrl Limit
  ------------ ---------- + --------- ------- ------ ---------- ---- ---- -----
  1/A18-Trk133 100/1000T  | No        Yes     Down   1000FDx    NA   off  0    
  1/A19-Trk133 100/1000T  | No        Yes     Down   1000FDx    NA   off  0    
  1/A20-Trk133 100/1000T  | No        Yes     Down   1000FDx    NA   off  0    
  2/A18-Trk133 100/1000T  | No        Yes     Down   1000FDx    Auto off  0    
  2/A19-Trk133 100/1000T  | No        Yes     Down   1000FDx    Auto off  0    
  2/A20-Trk133 100/1000T  | No        Yes     Down   1000FDx    Auto off  0   

When connecting a server via trunk e.g. to TRK133
the switch blocks the ports immediately. We've already changed all our attitudes
and the result is always without success.

We can't understand why
a) the trunks between the VSF core and access switches work and
b) the trunks between the VSF core switches do not work !!!!

[parnassus]

---

@AndyZ1 wrote:

We have a network with 2 Core HP switches configured as VSF.
Each switch has an identical design

...

We can't understand why
a) the trunks between the VSF core and access switches work and
b) the trunks between the VSF core switches do not work !!!!

---

Sorry, to me isn't clear: are you dealing with a pair of VSF (so 2+2 Aruba 5406R zl2) or are you dealing with only two Aruba 5406R zl2 deployed into one VSF?

Can you provide a clear image of your Network Topology? How VSF was setup (VSF Links)?

AFAIK, when you play with VSF exactly as when you play with a standalone Switch, both LACP Trunking configuration done against an Access switch and LACP Trunking configuration done against a physical Server is done on the VSF in the same way you do it on the Access Switch (permitted VLANs considerations apart)...clearly on the physical Server your NIC must support IEEE 802.3ad LACP bonding of its ports.

[AndyZ1]

Thank you for your quick reaction.
There are only 2 HP Aruba 5406 core switches.
The servers are HP DL380 Gen 8 servers and also have original HP network cards
Since LACP seems to work on the access switches (Aruba 2930F)
and not with the servers, we don't understand the situation.
Here the configuration of the two switches and setup of VSF.

; J9850A Configuration Editor; Created on release #KB.16.05.0004
; Ver #12:08.1d.fb.7f.bf.bb.ff.7c.59.fc.7b.ff.ff.fc.ff.ff.3f.ef:f6
hostname "SW-5400-001"
module 1/A type j9987a
module 1/B type j9987a
module 1/C type j9988a
module 1/D type j9987a
module 1/E type j9993a
module 1/F type j9993a
module 2/A type j9987a
module 2/B type j9987a
module 2/C type j9988a
module 2/E type j9993a
module 2/F type j9993a
vsf
   enable domain 100
   member 1
      type "J9850A" mac-address ecebb8-0fc800
      priority 150
      link 1 1/E8,1/F8
      link 1 name "I-Link1_1"
      exit
   member 2
      type "J9850A" mac-address ecebb8-d72500
      priority 100
      link 1 2/E8,2/F8
      link 1 name "I-Link2_1"
      exit
   port-speed 10g
   exit

show vsf link

 VSF Member 1
                 Link     Peer   Peer
 Link Link-Name  State    Member Link
 ---- ---------- -------- ------ ----
 1    I-Link1_1  Up       2      1   

 VSF Member 2
                 Link     Peer   Peer
 Link Link-Name  State    Member Link
 ---- ---------- -------- ------ ----
 1    I-Link2_1  Up       1      1   

show vsf link detail
VSF Member: 1     Link: 1
 Port     State                                                             
 -------- ------------
 1/E8     Up: Connected to port 2/E8                                        
 1/F8     Up: Connected to port 2/F8                                        

 VSF Member: 2     Link: 1
 Port     State                                                             
 -------- ------------
 2/E8     Up: Connected to port 1/E8                                        
 2/F8     Up: Connected to port 1/F8                                        

show vsf link utilization
               Link       Average Link
 Member  Link  Bandwidth  Utilization %
                           Rx        Tx
 ------  ----  ---------  -----     -----
 1       1     20G        3         0
 2       1     20G        0         3

show vsf topology
 VSF member's interconnection with links:
 Stby     Cmdr     
 +---+    +---+   
 | 2 |1==1| 1 |
 +---+    +---+   

show vsf topology detail
 Uni-destination Link Path:
 VSF Member ID  1    2   
 ------------- ---- ----
 1             -    1   
 2             1    -   

Multi-destination Path:
 VSF Member ID   Path
 -------------  ------
 1              2
   2                                      1

[AndyZ1]

Have you any idea to solve our problem !

[parnassus]

Hello! the VSF configuration looks good to me (I too have a pair of Aruba 5406R zl2 deployed as VSF).

What I don't understand instead is the status of your server facing trunk Trk133 ports (aggregated links).

Can you show us the output of commands listed below with regards to Trk133 (six interfaces [*] uplinked to your HPE ProLiant DL380 Gen8 server)?

• show lacp
• show lacp local
• show lacp peer
• show lacp counters

As I understood that Trk133 is untagged in VLAN Id 1.

I expect to see a IEEE 802.3ad LACP compatible configuration - server side - on those six NICs' ports.

Can you provide a proof of having done that too server side?

I would also consider to test the update to KB.16.05.0009 from your KB.16.05.0004 your VSF is running on...since there were corner case fixes related to LACP and SFP Transceivers that could have been eventually kicked in.

[*] Trk133 is done using six ports equally distributed on both VSF Members: 1/A18, 1/A19, 1/A20, 2/A18, 2/A19 and 2/A20.

[AndyZ1]

Good morning,

TRK133 is in VLAN1 (default) and the output of "show tech all" contains

vlan 1
   name "DEFAULT_VLAN"
   no untagged 1/A1,2/A1
   untagged
 1/A2-1/A12,1/C12-1/C24,1/D1-1/D24,1/E7,1/F1-1/F7,2/A2-2/A12,2/C12-2/C24,2/E7-2/
E8,2/F1-2/F7,Trk101-Trk124,Trk130-Trk135
   tagged Trk10-Trk26
   no ip address
   exit

LACP is configured on the server side.
We also have the strange situation that sporadically when a server is rebooted
the two TrunkPorts on the VSF Switch turn on and off and after a few seconds
minutes (see Appendix).
During this period, data backups are run and there is probably a
higher load.
Could the problem also be related to the VLANs?
Currently all servers are still running on the default VLAN VLAN1 !

Here the editions of the commands

show lacp
                                    LACP

           LACP      Trunk     Port                LACP      Admin   Oper
   Port    Enabled   Group     Status    Partner   Status    Key     Key
   -----   -------   -------   -------   -------   -------   ------  ------
   1/A13   Active    Trk130    Down      No        Success   0       1091  
   1/A14   Active    Trk131    Down      No        Success   0       1092  
   1/A15   Active    Trk131    Down      No        Success   0       1092  
   1/A16   Active    Trk131    Down      No        Success   0       1092  
   1/A17   Active    Trk132    Down      No        Success   0       1093  
   1/A18   Active    Trk133    Down      No        Success   0       1094  
   1/A19   Active    Trk133    Down      No        Success   0       1094  
   1/A20   Active    Trk133    Down      No        Success   0       1094  
   1/A21   Active    Trk134    Down      No        Success   0       1095  
   1/A22   Active    Trk135    Down      No        Success   0       1096  
   1/A23   Active    Trk135    Down      No        Success   0       1096  
   1/A24   Active    Trk135    Down      No        Success   0       1096  

   2/A13   Active    Trk130    Down      No        Success   0       1091  
   2/A14   Active    Trk131    Down      No        Success   0       1092  
   2/A15   Active    Trk131    Down      No        Success   0       1092  
   2/A16   Active    Trk131    Down      No        Success   0       1092  
   2/A17   Active    Trk132    Down      No        Success   0       1093  
   2/A18   Active    Trk133    Down      No        Success   0       1094  
   2/A19   Active    Trk133    Down      No        Success   0       1094  
   2/A20   Active    Trk133    Down      No        Success   0       1094  
   2/A21   Active    Trk134    Down      No        Success   0       1095  
   2/A22   Active    Trk135    Down      No        Success   0       1096  
   2/A23   Active    Trk135    Down      No        Success   0       1096  
   2/A24   Active    Trk135    Down      No        Success   0       1096  


show lacp peer
LACP Peer Information.
System ID: ecebb8-0fd83f

  Local  Local                          Port      Oper    LACP     Tx   
  Port   Trunk  System ID         Port  Priority  Key     Mode     Timer
  ------ ------ ----------------- ----- --------- ------- -------- -----
  1/A13  Trk130 000000-000000     13    0         0       Passive  Fast
  1/A14  Trk131 000000-000000     14    0         0       Passive  Fast
  1/A15  Trk131 000000-000000     15    0         0       Passive  Fast
  1/A16  Trk131 000000-000000     16    0         0       Passive  Fast
  1/A17  Trk132 000000-000000     17    0         0       Passive  Fast
  1/A18  Trk133 000000-000000     18    0         0       Passive  Fast
  1/A19  Trk133 000000-000000     19    0         0       Passive  Fast
  1/A20  Trk133 000000-000000     20    0         0       Passive  Fast
  1/A21  Trk134 000000-000000     21    0         0       Passive  Fast
  1/A22  Trk135 000000-000000     22    0         0       Passive  Fast
  1/A23  Trk135 000000-000000     23    0         0       Passive  Fast
  1/A24  Trk135 000000-000000     24    0         0       Passive  Fast

  2/A13  Trk130 000000-000000     205   0         0       Passive  Fast
  2/A14  Trk131 000000-000000     206   0         0       Passive  Fast
  2/A15  Trk131 000000-000000     207   0         0       Passive  Fast
  2/A16  Trk131 000000-000000     208   0         0       Passive  Fast
  2/A17  Trk132 000000-000000     209   0         0       Passive  Fast
  2/A18  Trk133 000000-000000     210   0         0       Passive  Fast
  2/A19  Trk133 000000-000000     211   0         0       Passive  Fast
  2/A20  Trk133 000000-000000     212   0         0       Passive  Fast
  2/A21  Trk134 000000-000000     213   0         0       Passive  Fast
  2/A22  Trk135 000000-000000     214   0         0       Passive  Fast
  2/A23  Trk135 000000-000000     215   0         0       Passive  Fast
  2/A24  Trk135 000000-000000     216   0         0       Passive  Fast
 

show lacp local
LACP Local Information.
               LACP                 Tx     Rx Timer                  
  Port  Trunk  Mode     Aggregated  Timer  Expired  System ID        
  ----- ------ -------- ----------- ------ -------- -----------------
  1/A13 Trk130 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A14 Trk131 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A15 Trk131 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A16 Trk131 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A17 Trk132 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A18 Trk133 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A19 Trk133 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A20 Trk133 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A21 Trk134 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A22 Trk135 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A23 Trk135 Active   Yes         Fast   No       ecebb8-0fd83f    
  1/A24 Trk135 Active   Yes         Fast   No       ecebb8-0fd83f    

  2/A13 Trk130 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A14 Trk131 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A15 Trk131 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A16 Trk131 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A17 Trk132 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A18 Trk133 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A19 Trk133 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A20 Trk133 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A21 Trk134 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A22 Trk135 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A23 Trk135 Active   Yes         Fast   No       ecebb8-0fd83f    
  2/A24 Trk135 Active   Yes         Fast   No       ecebb8-0fd83f    

[parnassus]

That 000000-000000 System ID on the output of show lacp peer is not encouraging...what is the configuration of aggreated ports, Server side?

I've many Port Trunks (Aggregated Links) to Servers on my VSF but, clearly, System ID is reported as known:

HP J9850A Switch 5406Rzl2
Software revision KB.16.05.0007

VSF# show lacp peer

LACP Peer Information.

System ID: f40343-2d103f

  Local  Local                          Port      Oper    LACP     Tx
  Port   Trunk  System ID         Port  Priority  Key     Mode     Timer
  ------ ------ ----------------- ----- --------- ------- -------- -----
  1/A23  Trk23  001185-d0f000     47    0         52      Active   Slow
  1/A24  Trk22  001185-d00040     47    0         52      Active   Slow
  1/B1   Trk100 000000-000000     33    0         0       Passive  Fast
  1/B12  Trk21  000af7-7341f6     7     128       48879   Active   Fast
  1/B13  Trk112 98be94-661876     3     128       48879   Active   Fast
  1/B14  Trk111 98be94-661832     3     128       48879   Active   Fast
  1/B15  Trk110 98be94-740dce     3     128       48879   Active   Fast
  1/B16  Trk109 98be94-740226     7     128       48879   Active   Fast
  1/B17  Trk108 000000-000000     49    0         0       Passive  Fast
  1/B18  Trk107 000000-000000     50    0         0       Passive  Fast
  1/B19  Trk106 000000-000000     51    0         0       Passive  Fast
  1/B20  Trk105 000000-000000     52    0         0       Passive  Fast
  1/B21  Trk104 000000-000000     53    0         0       Passive  Fast
  1/B22  Trk103 000000-000000     54    0         0       Passive  Fast
  1/B23  Trk102 e41f13-fbb0a5     1     128       48879   Active   Fast
  1/B24  Trk101 e41f13-fbb121     1     128       48879   Active   Fast
  1/C1   Trk1   00fd45-5c6800     1     0         962     Active   Slow
  2/A23  Trk23  001185-d0f000     48    0         52      Active   Slow
  2/A24  Trk22  001185-d00040     48    0         52      Active   Slow
  2/B1   Trk100 000000-000000     225   0         0       Passive  Fast
  2/B12  Trk21  000af7-7341f6     3     128       48879   Active   Fast
  2/B13  Trk112 98be94-661876     7     128       48879   Active   Fast
  2/B14  Trk111 98be94-661832     7     128       48879   Active   Fast
  2/B15  Trk110 98be94-740dce     7     128       48879   Active   Fast
  2/B16  Trk109 98be94-740226     3     128       48879   Active   Fast
  2/B17  Trk108 000000-000000     241   0         0       Passive  Fast
  2/B18  Trk107 000000-000000     242   0         0       Passive  Fast
  2/B19  Trk106 000000-000000     243   0         0       Passive  Fast
  2/B20  Trk105 000000-000000     244   0         0       Passive  Fast
  2/B21  Trk104 000000-000000     245   0         0       Passive  Fast
  2/B22  Trk103 000000-000000     246   0         0       Passive  Fast
  2/B23  Trk102 e41f13-fbb0a5     5     128       48879   Active   Fast
  2/B24  Trk101 e41f13-fbb121     5     128       48879   Active   Fast
  2/C1   Trk1   00fd45-5c6800     33    0         962     Active   Slow

reported System ID with 000000-000000 are unused / unconnected Port Trunks.

Trk1 (1/C1+2/C1) is used to connect to an Aruba 5412R zl2. Trunks in green have Servers NIC ports configured with LACP as peers.

[AndyZ1]

The servers are currently back on the old switch, because the system is used for production.
Therefore the ID is probably missing.

There is also a VMWare farm managed by an external service provider.
Each  of the 3 VMWare Servers DL380Gen8  have 8 network cards, 2 for management and 6 for the data network
(here  TRK131, TRK133, TRK135).
The customer assumes that the VMWare vSwitch is based on "iphash =source up IH hash". On Sunday we are on site, because the production does not work on Sundays.  
Should the V-Switch in the VMWare be set to "portid=Route based upon port id", so that the configuration in the VSG works with LACP trunk?

[parnassus]

Here a thread I partecipated with regards to vSphere vSS/vDS and LACP Port Trunking:

https://community.hpe.com/t5/Web-and-Unmanaged/HP-1810-24G-non-procurve-and-ESXI-vSphere/m-p/6992165

Notice that to use LACP you must deploy - or have yet - a vDS (Virtual Distributed Switch) vSphere side.

Usual vSS doesn't support LACP...so you're pretty much limited by VMWare.

Your Port Trunks are untagged into VLAN Id 1 so, as far as that is also replicated Server side...it should be OK.

Don't understand exactly the interlink you drawed on your network topology between 2530! Will it be a loop? considering they're uplinked to your VSF (at least if VLANs are the same between them).

In case of 2930F (stacked) there is not this issue since - I mean - the interlink is just the representation of the Stacking you probably did. Isn't it?