Power Cycle on POE ports with 1020 APs

[david.chinea]

We have 1020 access points with a combination of 4402 and 4136 controllers. We can watch the log in the 3750 switch and see them power cycle randomly. This, of course, is causing outages for the users. Has anyone seen this?

From our log:

May 18 10:38:14: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet2/0/42, changed state to down

May 18 10:38:15: %LINK-3-UPDOWN: Interface FastEthernet2/0/42, changed state to down

May 18 10:38:22: %LINK-3-UPDOWN: Interface FastEthernet2/0/42, changed state to up

May 18 10:38:23: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet2/0/42, changed state to up

May 18 10:38:27: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet2/0/42, changed state to down

May 18 10:38:29: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet2/0/42, changed state to up

[Dmitry Halavin]

This would indicate an AP reboot... check the management screen of the controller for AP logs, and check the message logs for AP resets ("last AP reset due to XXX"

[david.chinea]

I almost agree. The issue is that the down time is ~15 seconds. A typical reboot takes upwards of 60.

The controller doesn't log this as a reset. It doesn't log anything at all.

[mking]

Forgive the question, but are you on 3.2.116.21?

I noticed in the release notes that one of the items they fixed in the release was random reboots.