Symptom
*An increase in memory size for sessgmrd is observed in stand by switches wthe follwoing behavior has been observed:
*Every approximately every 2-3 weeks stand by switch in stack of Cat3k Switches reloads due a memory exhaustion.
*Core files are generated and logs point to a memory exhaustion issue as the cause of crash
*When this occurs , in stack of more than 2 switches, third switch in priority takes over the "stand-by" role, this switchover will not stop the leak and memory leak will be observed in new stand-by switch until the sessmgrd leak drains the memory from the new switch, once this occurs,a new switchover event will be triggerd, having stand-by switch role trasnferred back to the original stand-by switch
*Sessmgrd is related to dot1x, which is consistent with reports where environments that do not have dot1x implemented do not suffer to this issue
Conditions
The issue has been observed in stacks that are running stack and have dot1x implemented, devices with similar version that are not using Dot1x are not reported to be affected by this issues
Workaround
A reload can be performed on stand-by switch to restore memory utilization levels, however, this is a temporary measure as leak will persist.
Further Problem Description
The issue also affects Cat9K platform
This behaviour, stack members crashing, are NEVER observed on 3850/9300 running 16.12.4 with static VLAN assignments.
