Cisco Cisco 5760 Wireless LAN Controller Release Notes

Page of 34
 
32
Release Notes for Cisco 5760 Controller, Cisco IOS XE Release 3.2.xSE
OL-28114-04
  Caveats
CSCud68770
When you perform a continuous SNMPWALK on the table's attributes, the output is inconsistent.
When you perform a set on the cLD11ClientCalibTable, SNMPWALK gives the correct data for the 
first few minutes and then it does not return any data.
There is no workaround.
CSCud88714
When a nonhierarchical policy is installed on SSID output and when you try to overwrite it with a 
new policy which is in a hierarchical format, the policy change fails. This problem occurs only when 
a nonhierarchical policy is overwritten with a hierarchical policy.
The workaround is to unconfigure the existing policy and apply the new policy.
CSCud94109
If a client is roaming from Mobility Agent (MA) to Mobility Controller (MC) and joins another MA 
in a different peer group before complete authentication to MC, and then tries to rejoin to MC, the 
client entry cannot be deleted from the database. The client will not be able to join on the AP 
connected to MC but can join anywhere else in the network. 
The workaround is to use the test platform llm clear-database client_mac_address true command 
to remove the client entry on MC. 
CSCue44402
The controller displays the following message:
FRU Power Supply is not responding 
There is no workaround. 
CSCug23120
The show environment power all command randomly displays a power supply failure message and 
displays the wattage is displayed incorrectly as 235 W.
There is no workaround.
CSCug52183
When significant traffic (~ 4 billion packets) has traversed the CPU, the controller reloads 
unexpectedly. Depending on the control traffic pattern, it can take days or weeks for CPU-bound 
traffic to reach 4 billion. To check for this condition use the 
show platform punt stat port-asic 0 cpuq -1 direction rx command.
There is no workaround.
CSCug65693 
A Macbook client bug causes connectivity problems with a recent OS X update. This problem is 
triggered by the client sending an out of sequence packet. 
The workaround is to disable A-MPDU.
CSCug85580
When the auto qos voip cisco-phone command is applied to a port, data traffic over 10 (or 20) Mb/s 
is dropped at ingress ports.
The workaround is to remove the policer from the following class-map policy:
Service-policy input: AutoQos-4.0-CiscoPhone-Input-Policy
Class AutoQos-4.0-Default-Class
      set dscp default