Hi,
From my point of view, you have two options:
1. if the link need to be a l2 link you are stuck with Spanning Tree
2. If your link can be a L3 link, and I assume for the connection to your ISP it can, you can work with multi-path routing. I would create two transfer networks to the ISP router and would set different costs for the link (if not set automatically due to the different speeds). If the link with 1Gbps fails it should automatically use the route for the 100Mbps link.
------------------------------
-------------------------------------------------------------------------------
Florian Baaske
-------------------------------------------------------------------------------
Any opinions expressed here are solely my own and not necessarily that of Hewlett Packard Enterprise or Aruba Networks.
-------------------------------------------------------------------------------
Also visit the AirHeads Youtube Channel:
https://www.youtube.com/channel/UCFJCnuXFGfEbwEzfcgU_ERQ-------------------------------------------------------------------------------
Feel free to visit my personal Blog
https://www.flomain.de------------------------------
Original Message:
Sent: Aug 29, 2021 04:23 AM
From: MITF IT
Subject: LAG (LACP) 2 links different speed failover issue
Thanks, Davide
I understand that the Links should be same
but I am trying to use the LAG not for expanding the bandwidth, for my scenario the main reason for using it as Link failover
ٍSince I couldn't use LAG for my scenario could you advise me please if there is another thing I can use to make failover between two links (1000Mbps as the main ) (100Mbps as backup)
Thanks
Original Message:
Sent: Aug 28, 2021 08:45 AM
From: Davide Poletto
Subject: LAG (LACP) 2 links different speed failover issue
Hi Qais (MITF IT), a LAG - no matter if it is a Non-Protocol (called "Static") or LACP (called "Dynamic") - doesn't correctly form if media speeds of involved member physical links aren't identical (simplifying: LAG doesn't form if you aggregate a 100Mbps+1000Mbps or a 1000Mbps+100Mbps, LAG forms if you aggregate a 100Mbps+100Mbps or 1000Mbps+1000Mbps).
That's the very basic and essential rule to form a LAG and that is independent by Switch series, Vendor or OS used, it's a requirement due by design.
IIRC on various Link Aggregation Guides for Aruba CX Switches there is a statement about the above restriction: "Choose the same speed member interfaces while attaching to the LAG. Members with different speed than the already added member do not participate in aggregation for both static and dynamic LAG."
Also, another ESSENTIAL requirement is that the LAG's member links are co-terminating (ending) on the same physical (or logical, if any) switch where the peer LAG with same characteristic should exist...in other terms, you can't pretend a LAG perfectly formed on one device reach a good working state IF its member links aren't co-terminating on a peer device (physical or logical - as in case of virtual switch deployment - it is)...the way you aggregate two or more links together should be the same on both ends of the link.
Hope this clarify things a bit.
------------------------------
Davide Poletto