HPE 6125G switch using link-aggregation connect to Extreme Summit X450a-48t

[nnnnnnine]

Hi Everyone,

I got a trouble to connect 6125G to Extreme X450a-48t,

my 6125G firmware version is 5.20.99, Release 2106

below is my configuration sample:

[6125G]interface GigabitEthernet 1/1/5
[6125G-GigabitEthernet1/1/5]port link-aggregation group 17
[6125G-GigabitEthernet1/1/5]dis this
#
interface GigabitEthernet1/1/5
port link-aggregation group 17
#
return
[6125G-GigabitEthernet1/1/5]interface GigabitEthernet 2/1/5
[6125G-GigabitEthernet2/1/5]port link-aggregation group 17
[6125G-GigabitEthernet2/1/5]dis this
#
interface GigabitEthernet2/1/5
port link-aggregation group 17
#
return
[6125G-GigabitEthernet2/1/5]
 
[6125G]dis link-aggregation member-port GigabitEthernet 1/1/5
Flags: A -- LACP_Activity, B -- LACP_Timeout, C -- Aggregation,
       D -- Synchronization, E -- Collecting, F -- Distributing,
       G -- Defaulted, H -- Expired
 
GigabitEthernet1/1/5:
Aggregation Interface: Bridge-Aggregation17
Local:
    Port Number: 19
    Port Priority: 32768
    Oper-Key: 1
    Flag: {ACG}
Remote:
    System ID: 0x8000, 0000-0000-0000
    Port Number: 0
    Port Priority: 32768
    Oper-Key: 0
    Flag: {EF}
Received LACP Packets: 0 packet(s)
Illegal: 0 packet(s)
Sent LACP Packets: 0 packet(s)
[6125G]
 
 
[6125G]dis link-aggregation member-port GigabitEthernet 2/1/5
Flags: A -- LACP_Activity, B -- LACP_Timeout, C -- Aggregation,
       D -- Synchronization, E -- Collecting, F -- Distributing,
       G -- Defaulted, H -- Expired
 
GigabitEthernet2/1/5:
Aggregation Interface: Bridge-Aggregation17
Local:
    Port Number: 32
    Port Priority: 32768
    Oper-Key: 1
    Flag: {ACDEFG}
Remote:
    System ID: 0x8000, 0000-0000-0000
    Port Number: 0
    Port Priority: 32768
    Oper-Key: 0
    Flag: {DEF}
Received LACP Packets: 0 packet(s)
Illegal: 0 packet(s)
Sent LACP Packets: 11 packet(s)
[6125G]
 
[6125G]interface Bridge-Aggregation 17
[6125G-Bridge-Aggregation17]dis this
#
interface Bridge-Aggregation17
link-aggregation mode dynamic
mad enable
#
return
[6125G]dis link-aggregation verbose Bridge-Aggregation 17
Loadsharing Type: Shar -- Loadsharing, NonS -- Non-Loadsharing
Port Status: S -- Selected, U -- Unselected
Flags:  A -- LACP_Activity, B -- LACP_Timeout, C -- Aggregation,
        D -- Synchronization, E -- Collecting, F -- Distributing,
        G -- Defaulted, H -- Expired
 
Aggregation Interface: Bridge-Aggregation17
Aggregation Mode: Dynamic
Loadsharing Type: Shar
System ID: 0x8000, bcea-fa02-0219
Local:
  Port             Status  Priority Oper-Key  Flag
--------------------------------------------------------------------------------
  GE1/1/5          S       32768    1         {ACDEFG}
  GE2/1/5          U       32768    1         {ACG}
Remote:
  Actor            Partner Priority Oper-Key  SystemID               Flag
--------------------------------------------------------------------------------
  GE1/1/5          0       32768    0         0x8000, 0000-0000-0000 {DEF}
  GE2/1/5          0       32768    0         0x8000, 0000-0000-0000 {DEF}
[6125G]
%Apr 26 13:49:02:862 2000 6125G LAGG/5/LAGG_INACTIVE_PARTNER: Member port GigabitEthernet1/1/5 of aggregation group BAGG17 becomes INACTIVE because the port's partner is improper for being attached.
[6125G]

and on extreme side, i had done the following configuration:

enable sharing 1:1 grouping 1:1, 2:1 algorithm address-based L3_L4 lacp

And 6125G's G1/1/5 and G2/1/5 was connect to Extreme X450a-48t 1:1 and 2:1, 

When I finish my config. I will saw the messages from console like below :

%Apr 26 12:47:31:286 2000 6125G IFNET/3/LINK_UPDOWN: GigabitEthernet1/1/5 link status is UP.
%Apr 26 12:47:31:386 2000 6125G LAGG/5/LAGG_ACTIVE: Member port GigabitEthernet1/1/5 of aggregation group BAGG17 becomes ACTIVE.
%Apr 26 12:47:31:523 2000 6125G LAGG/5/LAGG_INACTIVE_PARTNER: Member port GigabitEthernet2/1/5 of aggregation group BAGG17 becomes INACTIVE because the port's partner is improper for being attached.dis interface GigabitEthernet brief

Can any one give me an advisor.

what kind of action should i take?

[parnassus · Accepted solution]

Why have you enabled the (LACP) MAD (Multi-Active Detection) on the LAG's configuration you have on your HPE 6125G?

Try to execute a (LAG-contextual) undo mad enable command to see if the LACP is finally negotiated correctly between the HPE 6125G and the Extreme Summit X450a-48t.

For sure both the Extreme and the 6125G should use not only the same LACP (That's OK) but also the same load-sharing algorithm (I see you used L3_L4 on the Extreme, see here).

On the HPE 6125G you can chose between a combination of the following criterias:

• Load-sharing criteria automatically determined based on the packet type
• Source IP address
• Destination IP address
• Source MAC address
• Destination MAC address
• Source IP address and destination IP address
• Source MAC address and destination MAC address

As example: link-aggregation load-sharing mode source-mac destination-mac

So you should cope the load sharing's mode on both your switches.

Side Note: consider to keep updated your 6125G Blade's Firmware (latest Firmware released was the R2112P04, March 2016...your one was released on July 2013). See here.

[Ian Vaughan]

Hello,

I've had something similar where the uplink cables from the Comware switch went into a 3rd party switch and were plugged into 2 ports in different LAGG groups at the other end of the link. Definitely worth checking. Other one to check (as ever) is for iffy / intermittent cable issues.

In other news:

I don't *think* that the MAD TLV's will break anything - If I remember correctly they just get silently filtered out by switches that don't understand them (rather than being reflected back as an IRF heartbeat by those that do).

The load sharing / hashing is an interesting one as sometimes (again IIRC) it can be necessary (or maybe even beneficial?) in some scenarios to have different configurations at each end of a link depending on how ancient your switches are.

If you have them available on the box, including higher layer protocols in the hash generally give a better spread - L4 good with range of negotiated TCP ports at the server end - not as good with single UDP port :-)

Please share your experiences and spread the word :-)

Ian