Hi there, first time posting :)
I'm setting up a few 4500x pairs in VSS and have been testing the fast hello feature included in the 3.5 IOS but have run into issues when I tested.
config for this portion:
switch virtual domain 5
dual-active detection fast-hello
interface TenGigabitEthernet1/1/14
description VSL FAST HELLO
switchport mode access
switchport nonegotiate
dual-active fast-hello
interface TenGigabitEthernet2/1/14
description VSL FAST HELLO
switchport mode access
switchport nonegotiate
dual-active fast-hello
We cabled up 4 copper VSLs and 1 copper fast hello port. then we setup a switch via etherchannel, fed to both 4500x. we setup pings to that switch as well as the 4500x pair. then we dropped all 4 VSL links. now correct me if i'm wrong, but at that point we should see one of the chassis reboot and dual-active detection should respond with a YES to the following command:
test#show switch virtual dual-active summary
Executing the command on VSS member switch role = VSS Active, id = 1
Pagp dual-active detection enabled: Yes
FastHello dual-active detection enabled: Yes
In dual-active recovery mode: No
Executing the command on VSS member switch role = VSS Standby, id = 2
Pagp dual-active detection enabled: Yes
FastHello dual-active detection enabled: Yes
In dual-active recovery mode: No
Well that's not what happened. What did happen is the standby chassis went all amber, on all links. From the active, i saw everything drop except my link to my core. The fast-hello port went down, even though both chassis were still powered up (the standby did NOT reboot as i expected). At this point I tried consolling into the standby with the amber on all ports...no go. Could not get a prompt of any kind. I then plugged one of the VSLs back in and saw the following in the logs:
Dec 4 14:03:44.473 EST: %C4K_IOSINTF-5-LMPHWSESSIONSTATE: Lmp HW session UP on slot 1 port 11. ***typically see when one is on and the other one is rebooting fresh
Dec 4 14:04:00.472 EST: %VSLP-5-VSL_UP: Ready for control traffic ***typically see when one is on and the other one is rebooting fresh
Dec 4 14:04:04.476 EST: %VSLP-5-RRP_ROLE_RESOLVED: Role resolved as ACTIVE by VSLP ***looks ok so far
Dec 4 14:04:04.477 EST: %EC-5-BUNDLE: Interface TenGigabitEthernet1/1/11 joined port-channel Port-channel5 ***looks ok so far...
Dec 4 14:04:07.305 EST: %VSLP-3-VSLP_LMP_FAIL_REASON: Te1/1/11: Link down ****uh oh
Dec 4 14:04:07.305 EST: %VSLP-2-VSL_DOWN: All VSL links went down while switch is in ACTIVE role ****and goodbye to that switch
So the last two lines show the one active VSL dying for some reason. It's at this point the switch in amber reboots completely, even though it has a valid VSL link to the active switch again. When it reboots, everything is fine again. At no point during all of this did the "show switch virtual dual-active summary" command or fasthello show command show that it had detected the dual active. This is not a surprise as the port was physically down (even though still physically plugged in with both chassis powered).
I'm figuring it's a bug with this new IOS feature but I wanted to check with everyone in case I was doing something wrong or someone had run into this before. We tested this twice in a row with the exact results.
Oh, and I never lost a ping to the switch on the downstream of these two switches, so that's good at least. No loops or anything that I could tell.
