I had a whale of a Saturday in in my COLO today trying (and failing) to complete what should have been a very simple task. I still do not understand why this happened, or how to fix it, but here it is.
Stacked Force10 MXL switches inside an M1000e chassis with VMware ESXi 5.5 cluster on the inside backed by the Equallogis M4001x storage set.
3 Blade servers inside all running NPAR so all inernal links are tagged.
I have been running 4 vlans forever: 1.Internal networking for virtual hosts. 2.ISCSI 3.Vmotion 4.External connectivity with other hosts on outside LAN subnets. This is fine, and dandy. I need to set up some connectivity to the Internet now, and Have added a 5th VLAN to my stack, and tagged the ports appropriately in bring the connectivity "inside" to the VM that needs the external connection. Simple. Now:
VLAN 10. Is Vmotion
VLAN 20 is iSCSI
VLAN 100 is my internal LAN
VLAN 200 is external LAN subnets
VLAN 300 is Internet,
Here it comes....when I plug in the cable (VLAN300) between my Internet switch, and the MXL the Internet comes up..HOORAY!
AND....(VLAN200) is unreachable. Now I can SSH into my M1000e, but cannot reach it from inside the organization.
Until - I unplug the cable leading to VLAN200 and then plug it back in then...VLAN200 becomes reachable, but VLAN300 is now inaccessable.
I tested this 9 ways to Saturday night today, and this is all I have learned. I cannot find any logging on the MXL stack that will give me a clue as to why this is happening, nor can I explain this myself. I have ancient HP switches that can handle this kind of connectivity, but these cutting edge interconnects from Force10 cannot? What gives?
Is there a maxvlans setting on these things that just will not allow those two external/internal vlans to exist together?
Anyone seen anything like this? I plan on calling DELL support about this first thing Monday if I can't get any help here. Thanks for any comments you leave me.
Solved! Go to Solution.
Perhaps I need to add more color. I will post a picture in the morning of what is happening, and open a case with Force10 support.
A call into support would be a good idea. Depending on the type of contract you have, they can remote in and help with configuration.
If you have a topology diagram you can post up with your running config. It would help us get a better understanding of setup and what might be configured incorrectly.
OK. I got the picture uploaded, and you can see it above. I will upload the running config next, but it is rather long as these switches have 110 interfaces. Is there a way to get this to you without copy/paste it all in here?
Thank you Daniel. You led me to STP as the culprit and you were correct. It turned out that my COLO was running STP on their side, and that was causing my MXL to block one of the other of my ports. I did some extensive re-routing of cables on my side of things and removed all possible loops, and things worked. I also