Core issue
This is a rare out-of-synchronization condition on the fabric-channel between a particular module and the switch fabric on the supervisor engine which can lead to affected module power cycle, permanent power down or supervisor switchover in dual supervisor configurations and cause partial network outage.
This issue is due to false classification of fabric-channel events by switch fabric firmware triggering excessive fabric-channel resynchronization. Excessive resynchronization events can also contribute to the increase of rxErrors reported by the receipt of the fabric-enabled module, which is documented in Cisco bug ID CSCsc55949.
Resolution
In order to take advantage of the redesigned fabric-channel monitoring mechanism, the recommended action is to upgrade to Cisco IOS Software Release 12.2(18)SXE5 or Cisco IOS Software Release 12.2(18)SXF3 or later, which can be downloaded from Cisco Downloads.
When you run the Cisco IOS software releases prior to 12.2(18)SXF3, the workaround is to issue this command in order to minimize the risk of this problem:
sup720#remote command switch test fabric flap count 10
sup720#remote command switch test fabric flap time 1
