Loading...
Loading...
3850 with low memory due to memory leak in NGWC SPI Async Response here the scenario is dot1x/mab authentication is enabled, there are lot of authentication failures, link flap with radius server, with sessions coming in and going out, leading to memory leak. It can take box one day to week to go out of memory depending on number of sessions coming in.
Dot1x/mab authentication is enabled.
These are the known workaround, On 3.3.4SE release. Populate credentials of all devices in radius server. Ensure radius connectivity. Monitor memory usage, if it is going beyond 70%, reset the box. On 3.6.1 If customer is having config "authentication event fail action next-method? then no memory leak is seen. On 3.6.1 if the customer is having the config "authen event fail action authorize vlan ? then by disabling 'aaa accounting' memory leak was no longer seen. no aaa accounting dot1x default start-stop group radius no aaa accounting network default start-stop group radius
Cisco Integration
Learn more about where this data comes from
BugZero Plan
Streamline upgrades with automated vendor bug scrubs
BugZero Prevent
Wish you caught this bug sooner? Get proactive today.