8 replies to this topic

[Rick Young]

Exception running Task: Task execute failed: Returned Request ID has unexpected value

The above error was retrieved from Java Web Service's log file. Anyone know what the latter part means? (Returned Request ID has unexpected value)

There was no action taken to rectify - this is a task/function that has existed for a while, and in fact got called again successfully about 10 seconds later.

The task that was running is the most simplistic: no incoming parms, 2 out going parms of an error code and error message. Simply returns an error code of 00 and an "all good" message - simply to test if the server is available.

-Rick

[Daniel Elkington]

Are you using v5.5 or v6? I've come across that same error message caused by a wide variety of issues, the error reporting on the 5.5 Java bus really is pretty useless. Was told at the time that this should be improved in v6 but i don't know if that's actually happened.

[Rick Young]

Remiss of me not to include that information.

ProIV server is HP-UX 11i 64-bit, v5.5r517 32-bit.
Customer is not running Concerto, per se, just has the Concerto-licensed .jar files implemented in Java Web Services running on a Win2k3 SP1 machine, with the Tasks being invoked from C#.NET app in IIS on the same physical Win2k3 box.

ProIV support (sorry, failed to mention herein my intent to contact them also), did reply initially with this error possibly being either a) network congestion to the Concerto license [seems improbable in my scenario]; some "hang" in the pro4busd requiring restart of the daemon [which was not required in my scenario] - and perhaps related to my other topic suspecting pro4busd memory leak.

They did further suggest (and provided code) to turn on verbose debug logging in the JWS, however, that is going to require some "committee discussion" before that event occurs on my customer's Production box...

-Rick

[Daniel Elkington]

Just to clarify, do you mean you're using the java API, or the java bus, or both? We have a customer running the java bus, and they had what we initially thought were licencing problems, but turned out to be caused by heavy network traffic and server load, these were fixed by tweaking the settings in the config.xml file that controls the bus parameters.

[Rick Young]

In order for me to answer your question, could you define for me the java bus?

We have no config.xml file.



-Rick

[Daniel Elkington]

You've answered my question then, if you don't have a config.xml file then you're not using the java bus. The config.xml file controls all the settings like number of simultaneous connections, timeouts for starting new kernels, etc. I haven't used the standard bus much, so i don't know if you can control those parameters some other way.
In the situation i was referring to before, the timeout limit for starting a new kernel was too low, which meant that when the server was busy and a task call was made, the server was taking too long to start up a new kernel process. I think there are significant differences between the standard bus and the java version though, so i'm not sure whether that situation would be applicable.

[Chris Mackenzie]

Exception running Task: Task execute failed: Returned Request ID has unexpected value

The above error was retrieved from Java Web Service's log file. Anyone know what the latter part means? (Returned Request ID has unexpected value)

TCP/IP error. But have you got the API error code?

Rather than enabling verbose logging, do you have event logging
enabled?

P4EVENTREPORTLEVEL=4

in pro4.ini bus section values go from 0 to 4. If err occurs again check
the syslog.

[Rick Young]

Thanks for responses guys.

Daniel: No timeout on starting kernels by the pro4busd on the HP-UX box. Succesfully have thrown 5 simultaneous requests (max kernels set to 5), and have "successfully" thrown a 100 requests (with the resulting expected license seat failures).

Chris: turning on verbose logging (and/or tracing) on the Production box is not real high on the desire list...I filled a drive recently...say no more. I'm puzzled with the combination of your comments with TCP/IP error, and wanting the API error code. The Java Web Services has is not on a ProIV box, and is not within Concerto. If you're indicating a comms issue between the JWS Win2k3 box and the HP-UX box, then pro4-side logging I wouldn't anticpate to provide anything(??) If the comms issue is internal of the Win2k3 box between IIS and JWS, and we don't have the config.xml from Concerto and/or Java Bus, then there's nothing to log there either - except by perhaps turning on Java logging.



For now though, thanks for your input and we shall re-address should it occur again and/or if my customer agrees to turn on logging/tracing.

Cheers
Rick

[Chris Mackenzie]

Thanks for responses guys.

Chris: turning on verbose logging (and/or tracing) on the Production box is not real high on the desire list...I filled a drive recently...say no more. I'm puzzled with the combination of your comments with TCP/IP error, and wanting the API error code.

Hi,

If you get the API return code it will prove if the connection was made and broken
or never made. Maybe not very useful info...

Enablling event logging in the bus daemons doesn't write nearly as much
data as enabling tracing - but as you say, if you're not connecting at all then
it won't be very useful anyway.