Ethernet Switching
Highlighted
Ethernet Switching

LAG Issue between Juniper EX2200-C and Cisco 3560 Layer 2 loop

[ Edited ]
‎03-18-2018 07:13 AM

Having some issues with a very simple Lab scenario between an EX2200-C and a Cat 3560. Both devices are basically in default configuration. The Juniper has the configuration listed below and the topology is listed below that. My issue is that when I have the Catalyst with no LAG/Etherchannel configuration I end up getting a Layer 2 loop. It seems that the member interfaces on the Juniper are still forwarding traffic even though the Aggregated interface is down. I'm used to seeing the Cisco devices put the interfaces into an individual state and keep them running in STP to prevent loops.

 

Is this behavior normal?

 

----
Notes 1: This is a lab test environment
Notes 2: You shouldn't have one end not configured for LAG but this is a test to see what happens in that case
Notes 3: I'm trying to simulate the scenario where one party provisions a LAG before the other
EX2200-C is running: 12.3R12 ############################################################################ Spanning tree interface parameters for instance 0 Interface Port ID Designated Designated Port State Role port ID bridge ID Cost ge-0/1/1.0 128:562 128:562 32768.3c61046021c1 20000 FWD DESG {master:0} ############################################################################ root> show configuration chassis aggregated-devices { ethernet { device-count 1; } } root# show interfaces ge-0/0/0 ether-options { speed { 100m; } 802.3ad ae0; } root# show interfaces ge-0/0/01 ether-options { speed { 100m; } 802.3ad ae0; } root# show interfaces ae0 description "Link to DSW1"; aggregated-ether-options { lacp { active; } } unit 0 { family ethernet-switching; } ############################################################################ LAG +--+ | | +---------------------------------------------+ |---------------------------------------------| || | | || || +--+ || || || ge-0/0/0||ge-0/0/1 ||f0/1 +--------------+ +--------------+ | | | | | EX-2200C |ge-0/1/2 g0/1| c3560 | | +-------------------------+ | | | | | +--------------+ +--------------+   
#############################################################################

 

3 REPLIES 3
Highlighted
Ethernet Switching

Re: LAG Issue between Juniper EX2200-C and Cisco 3560 Layer 2 loop

‎08-09-2018 04:17 AM

I would suggest this behavior is normal and expected.  Since LAG is a local implementation, that is there is no standard covering this technology, there is no way for a LAG on one end to know there is non-LAG on the remote side.  I am thinking if you enable some form of STP on both sides, this may cause non-LAG side to block one link. 

 

Juniper defaults STP to type RSTP and you would also need to set VLAN 1 as default-vlan, as unlike Cisco this is not the default behavior for Juniper.  This assumes VLAN 1 is not the only VLAN assigned to this LAG.

 

I do assume you have LACP active on both sides, but this operates at per link level, not LAG level.

 

Good luck, HTH.

Highlighted
Ethernet Switching

Re: LAG Issue between Juniper EX2200-C and Cisco 3560 Layer 2 loop

‎08-09-2018 04:54 AM

So LACP only has the role of bundling/removing the interfaces into the aggregated-ethernet interface, as long as there is no LACP and LAG it's up to the switch what it does with those unbundled interfaces?

Highlighted
Ethernet Switching

Re: LAG Issue between Juniper EX2200-C and Cisco 3560 Layer 2 loop

‎08-09-2018 05:20 AM

LACP is really completely independent of LAG, although it is most often used with LAGs.  You can run LACP on a single link, or a LAG of one.  LACP is really there to protect against one-way faults, which would cause one end to think their link is up when it actually is not.  In a LAG of multiple interfaces, this creates a potential black-hole.  Before LACP Bi-Directional Forward Protection was often used, but not a standard.

 

LACP will not determine if either side is configured in a LAG.  Functions on a per-link basis, from what I know!

 

HTH