Cisco Cisco IOS Software Release 12.4(4)T

CSCsb15138

Symptoms: The following error messages may be generated on a gateway that functions in a 
configuration in which 80 channels are processed by a VXML Server, and the call may be dropped:

//-1//HTTPC:/httpc_streaming_create: attempt to create a session with id 699 

while this id is in use

//2144684/0BCEFBA9AA28/VXML:/vxml_media_done: 

CALL_ERROR; fail with vapp error 2, protocol_status_code=0

//2144684/0BCEFBA9AA28/VXML:/vxml_media_done: 

CALL_ERROR; *** error.badfetch.http.0 event is thrown

Conditions: This symptom is observed rather rarely on a Cisco AS5400 gateway when the HTTP 
client session IDs range from 1 to 2048 because of the socket limit per Cisco IOS process. The error 
messages are generated when the HTTP client attempts to create a new session with the same ID as 
an old session that is still in use. In this situation, only a benign warning message should be 
generated, and the call should be accepted. If an HTTP streaming session remains in use for a long 
time and the traffic load of the gateway is high, the symptom is more likely to occur.

Workaround: Configure an event handler as in the following example:

<catch event="error.badfetch.http.0">

<!--  Actual event handler goes in here -->

</catch>

If this is not an option, the symptom may be mitigated by disabling IVR streaming mode via the ivr 
prompt streamed none command. 

CSCsc33783

Symptoms: When using Object Tracking in Cisco IOS, some configuration may be lost on reload 
depending on the configuration of the object tracking references. Objects are parsed in sequential 
order so if one tracked object references a second tracked object of higher number then that second 
object will not be defined when the first object is initialized on reload.

Conditions: This symptom occurs when using Object Tracking in Cisco IOS.

Workaround: Always use a higher numeric object tracking ID for the parent object. 

CSCsd28214

Symptoms: A Cisco router that is running Cisco IOS Release 12.3(19) may crash due to a Watch 
Dog timeout while running the RIP routing protocol.

Conditions: The router may crash due to a Watch Dog timeout if an interface changes state at the 
exact same time a RIP route learned on that interface is being replaced with a better metric 
redistributed route. For example, RIP has learned the 192.168.1.0 network from Fast Ethernet 1/0. 
If RIP learns the 192.168.1.0 network from a redistributed protocol that has a better metric, then the 
RIP route will be removed. If, during this time the Fast Ethernet 1/0 interface goes down, then the 
router may potentially crash due to a Watch Dog timeout.

Workaround: There is no workaround.