Apache TimeOut and system (pxplus) not terminating

[yonman]

As per documentation, we've a default timeout of 30 seconds for an Apache web service installation.  It's not been overridden for years.  Starting last year, we've noticed that pxplus sessions, periodically, do not terminate.  They don't seem to use any pxplus web licenses while they are hanging there.  But, it does slow down the performance hugely, however.

The hug sessions are a variety of web pages, not any one specific page. 

Might be a port issue.  But, I thought this might be an easy answer for the Pxplus Developers here.

Thank you for any suggestions!
(attached)


timeout= nn

Changes the default 30 second timeout for the maximum run time that a task can take. If a spawned task takes longer than the number of seconds in the timeout period, the system will terminate the task and report an error.

[EVa]

The PxPlus Apache interface runs the *plus/apache/pxp_cgi program which relies on the *SYSTEM SetTimer() function to force a timeout (https://manual.pvxplus.com/?utilities/SYSTEM%20Event%20Handling%20Object/Introduction.htm).  I'm assuming that your web application is not resetting that timer as that would cause problems with the Apache timeout.  Did you try to connect to these sessions using the 'dbg' console mode command to determine where in your application they are stuck (https://manual.pvxplus.com/?commands/dbg.htm) ?

[yonman]

Hi Eva,

Thank you for those two thoughts.  The DBG does not connect to these sessions, however I am able to identify what html pages they are viewing.  So, I'll investigate your thought about the SetTime function.

=) Thank you.

Best,
Yonman

[yonman]

Just a quick update on this problem:

1.  dbg cannot connect into this running pvx web session.  However, the cpu cycle time shows it's running.
2.  I was able to trace the pvx web session on the client's SQL server and found this being reported in Activity Monitor Session:

-- network protocol: TCP/IP
set quoted_identifier on
set arithabort off
set numeric_roundabort off
set ansi_warnings on
set ansi_padding on
set ansi_nulls on
set concat_null_yields_null on
set cursor_close_on_commit off
set implicit_transactions off
set language us_english
set dateformat mdy
set datefirst 7
set transaction isolation level read committed

...suggesting a bad handshake?