We are a small office with two Dell PowerConnect 5548 switches that, together, cover all of our servers and hosts. I had them connected using ports 47 and 48, which I placed in a LAG. Port 47 was using a fiber converter box on each end that was only capable of 100Mbps. For port 48, I was able to gain access to a CAT6 cable that spanned the two rooms (approx 150ft apart). Some backups I was doing were getting capped at 100Mbps, so I reasoned that the LAG was just choosing one of the links and using the other as failover. To test that out I unplugged port 47, thinking port 48 would kick in and we'd get Gig speeds. Instead, all communication between the two switches was cut off. Thanking that spanning tree timers might be involved somehow, I let it sit for a little bit (2 minutes?) to see if it would fix it self, but it did not. So I plugged port 47 back up, and things came back up. As a temporary "fix," I have replace the fiber link with a CAT6 all the way so as to avoid the 100Mbps bottleneck.
My question is certainly reflect my lack of knowledge, but I'm okay with that. I only ever use the GUI. That's how I set up the LAG. So some things I can see by looking at stuff in there: The LAG (#1) is set as a trunk. Its PVID is 1. The Frame Type is "Admit All." Ingress Filtering is set to Enable. Regarding ports 47 and 48, Auto Negotiation is Enabled, Back Pressure is Disabled, Flow Control is Enabled, and LAG number shows as 1 for both. However, there is a difference between the two: port 47 has an Auto MDIX setting of "MDIX," whereas port 48 has a setting of "MDI." Not sure if that matters.
Why didn't the failover work? Is something configured improperly? Am I misunderstanding how a LAG should work?
