...
Stack-Wise Test on e Catalyst9300 :During the simulation of dual stack cable failure, isolate the Active member and create a dual brain scenario with two stacks each of them with an Active switch (as below): Stack 1: Standalone Cat-9300#show switch Switch/Stack Mac Address : cc5a.53d9.9a80 - Foreign Mac Address Mac persistency wait time: Indefinite H/W Current Switch# Role Mac Address Priority Version State ------------------------------------------------------------------------------------- 1 Member 0000.0000.0000 0 V01 Removed 2 Member 0000.0000.0000 0 V01 Removed *3 Active cc5a.53d9.b180 1 V01 Ready Cat-9300#show switch stack-port Switch# Port1 Port2 ---------------------------- 1 DOWN DOWN 2 DOWN DOWN 3 DOWN DOWN Cat-9300#show switch stack-ring speed Stack Ring Speed : 240G Stack Ring Configuration: Down Stack Ring Protocol : StackWise HAGKK4-LAB-03AS02# Stack 2: Partitioned stack Cat-9300#show switch Switch/Stack Mac Address : cc5a.53d9.9a80 - Local Mac Address Mac persistency wait time: Indefinite H/W Current Switch# Role Mac Address Priority Version State ------------------------------------------------------------------------------------- 1 Standby cc5a.53d9.9a80 1 V01 Ready *2 Active cc8e.714b.bc00 1 V01 Ready 3 Member 0000.0000.0000 0 V01 Removed Cat-9300#show switch stack-ports Switch# Port1 Port2 ---------------------------- 1 DOWN OK 2 OK DOWN 3 DOWN DOWN When we bring back the two Stack Cables , the portioned stack is reloaded and during the stack merge we observe that the Standby Switch get stacked in HA sync state (as below). Only a reload of the entire stack brings everything to the right state. HAGKK4-LAB-03AS02# show switch ---------------------------------------------------------- 1 Standby cc5a.53d9.9a80 1 V01 HA sync in progress 2 Member cc8e.714b.bc00 1 V01 Ready *3 Active cc5a.53d9.b180 1 V01 Ready
Stack-merge
1) Rejoin the stack cables one by one giving some seconds gap or 2) Issue a reload of the stack
Observed also in Cat3650.