When does the clock start for the request? Does it start when the reactor thread receives the request?

Document ID : KB000012491
Last Modified Date : 14/02/2018
Show Technical Document Details
Introduction:

This question/answer below relates to the fact that the trace log can  show a messages like CSm_Auth_Message::AnalyzeAgentAuthMessage that rarely take a long time to  complete.  

At the same time you received a TCP timed out message, the CSm_Auth_Message::AnalyzeAgentAuthMessage  took a long time to  complete in this type of scenario.

The TCP time out is not the cause of CSm_Auth_Message::AnalyzeAgentAuthMessage  taking  a long time to complete. It is just reflection of the CSm_Auth_Message::AnalyzeAgentAuthMessage  took a long time to  complete.

This question/answer -  answers  one of  part of the flow when this type of scenario happens.  

Question:

When does the TCP timeout or any clock start for the request? Does it start when the reactor thread receives the request? 

Answer:

The TCP timeout clock or any clock should begin with the request. The time stamp is in the request . Not when the reactor thread receives the request

Additional Information:

A pool of worker threads is utilized to handle requests from the Web Agent. A reactor thread handles the initial connection from all incoming requests and passes them into a queue serviced by the worker threads. The position in the queue can differ depending on the type of incoming request 

The web agent should time out earlier than message in the queue would and the agent would retransmit the message - possibly to another policy server 

If the normal priority messages timed out in the Policy Server queue, the [TCP Message timed out. It will be discarded.] 

Message is logged and the request is discarded from the queue. This TCP timeout is governed by the RequestTimeout parameters defined in the HCO

 

The RequestTimeout parameters, which are set in the Host Configuration file (HCO), SmHost.conf.

Settings in the Host Configuration File apply only when the trusted host initializes. Once

the trusted host initializes, the settings in the Host Configuration Object take effect.

Request Timeout

Use the RequestTimeout parameter to specify the number of seconds that the trusted

host should wait before deciding that a Policy Server is unavailable. This setting allows

you to optimize the response time of the Web server.

The default value is 60 seconds.

Note: If the Policy Server is busy due to heavy traffic or a slow network connection, you

may want to increase the RequestTimeout value.