Intermittently receiving 404 page not found when accessing the Web based Application.
Traces shows requests going to back-end on wrong port / going to wrong back-end, which responds with 404 as it does not have the resource being reuested

In the snapshot below,

Request for 10.107.189.115/8080/page1.html should be fwd to 10.107.189.114:8080/page1.html 
Request for 10.107.189.115/9090/page2.html should be fwd to 10.107.189.114:9090/page2.html 

However intermittently we see:
Request for 10.107.189.115/8080/page1.html fwd to 10.107.189.114:9090/page1.html 
Request for 10.107.189.115/9090/page2.html fwd to 10.107.189.114:8080/page2.html 

Both these requests will result in 404 error from server as 10.107.189.114:9090 is not hosting page1.html and 10.107.189.114:8080 is not hosting page2.html 




 

Recommended - Disable "Connection multiplexing" on the http profile bound to  wildcard LB VIP or wildcard Servcie / Service Groups.
Or do the same globally (nshttp_default_profile) - note this will affect all VIPs

---

Problem Cause

This behavior is due to  connection multiplexing being enabled on the VIP.  With wildcard service groups and VIPs connection multiplexing can cause these issues as Netscaler tries to reuse existing open connections for multiple http requests.