We have several jobs running under 24x7. We are currently having a problem where by some jobs just stop running.
This one in particular runs a dos batch file.
Other jobs seem to work alright but at odd moments the job just stop working.

On attempting to close 24x7 it gives me an error that a spawn process is still running,shutting down may cause unpredictable results.
It then refuses to shutdown when I select shut down anyway.

We have attempted recreating jobs like these and they do work for a period before they exhibit similar problems.

Job stop running because they might be blocked by a "runaway" job. Use the Job Queue monitor to find out which job(s) causes the trouble. When you find that job assign it to a separate queue. Create additional job queues as needed.

: We have several jobs running under 24x7. We are currently having a problem
: where by some jobs just stop running.
: This one in particular runs a dos batch file.
: Other jobs seem to work alright but at odd moments the job just stop working.

: On attempting to close 24x7 it gives me an error that a spawn process is
: still running,shutting down may cause unpredictable results.
: It then refuses to shutdown when I select shut down anyway.

: We have attempted recreating jobs like these and they do work for a period
: before they exhibit similar problems.

: Job stop running because they might be blocked by a "runaway" job.
: Use the Job Queue monitor to find out which job(s) causes the trouble.
: When you find that job assign it to a separate queue. Create additional
: job queues as needed.

This is not the case. The program exist in a job queue by itself. The size of the queue is 2MB.

Are there any other possible solutions?

If a job hangs in the queue it will block itself (all consecutive runs) and all other jobs that are assigned to the same queue.
This is exactly the situation with a "runaway" job.

Please tell us more about the job so I can suggest how to deal with the problem.

PS. Size of the queue has nothing to do with the described problem.

: This is not the case. The program exist in a job queue by itself. The size of
: the queue is 2MB.

: Are there any other possible solutions?

Hi.

I'm working along with Andre - the batch job runs
a C++ program which accepts 3 parameters and then
executes an 'EXIT' command after the termination
of the C++ program.

I've now set the job to run detached and am watching
it closely now.

Is there anything else you need to know?

Camille Moore

: If a job hangs in the queue it will block itself (all consecutive runs) and
: all other jobs that are assigned to the same queue.
: This is exactly the situation with a "runaway" job.

: Please tell us more about the job so I can suggest how to deal with the
: problem.

: PS. Size of the queue has nothing to do with the described problem.

Try this, create a batch file which will run your program and schedule this batch instead of main EXE.

In the batch file code

echo Stated
your_EXE_command_line
echo Finished

Schedule this batch file as
batch_file_name.bat >> batch_file_name.log

This way you can verify if your EXE is causing the problem or it is something else.
If the log file after your EXE run does not have "Finished" message, your EXE is still running.

Run your job for a while to find out if the EXE always terminates.
Hope this will help you to find out why the job hangs in the queue.

Also consider using the timeout option to let 24x7 forcibly terminate the EXE (or batch) if it takes too long to run.

: Hi.

: I'm working along with Andre - the batch job runs
: a C++ program which accepts 3 parameters and then
: executes an 'EXIT' command after the termination
: of the C++ program.

: I've now set the job to run detached and am watching
: it closely now.

: Is there anything else you need to know?

: Camille Moore

Ok - our batch job is already set up this way - the
only difference being the echo statements which I
will add now and watch for a few hours.

We do in fact have a 3 minute timeout on the job,
and based on what I see in the log file, this 3
minute limit is never reached.

Will keep you posted.

Regards.

: Try this, create a batch file which will run your program and schedule this
: batch instead of main EXE.

: In the batch file code

: echo Stated
: your_EXE_command_line
: echo Finished

: Schedule this batch file as
: batch_file_name.bat >> batch_file_name.log

: This way you can verify if your EXE is causing the problem or it is something
: else.
: If the log file after your EXE run does not have "Finished"
: message, your EXE is still running.

: Run your job for a while to find out if the EXE always terminates.
: Hope this will help you to find out why the job hangs in the queue.

: Also consider using the timeout option to let 24x7 forcibly terminate the EXE
: (or batch) if it takes too long to run.

Hi again:

We had problems with our scheduled job again last night - please see below for the report I had to give to my manager:

At 1:18am, the payment loader module timed out. There was no reason given, but there were the following lines in the log file:

11/19/2002 01:15:50 0 0 24x7 Scheduler 24x7 Scheduler starting...
11/19/2002 01:15:51 0 210 Payment Loader Job started.
11/19/2002 01:17:26 0 197 Check CRM Job started.
11/19/2002 01:17:27 0 197 Check CRM Job finished.
11/19/2002 01:18:51 1 210 Payment Loader Timeout. Process 'Z:\Apps\AWBS\Script\PayLoad.bat >> z:\Apps\AWBS\output\PayLoad.log' has been terminated!

Thereafter, the payment loader ran every 5 minutes as scheduled until 3:20am either reporting an exit code of 1 each time or not completing at all; i.e. at 3:00am, it was started but did not seem to complete since there is no accompanying "finished" message.

At 3:30am, the Package Loader was reported to have been started also but there was no report of completion though the package loader seemed to running to completion thereafter.

These two "uncompleted" processes would be the cause of the "11/19/2002 07:07:25 1 0 24x7 Scheduler 2 spawned processes are still running. Shutting down the scheduler now may cause unpredictable results." message this morning when a shutdown of the scheduler was attempted.

The checks I put into the PayLoad.bat file itself did not yield any further information and I have since modified the checks so that time and date information is also recorded in a log file.

Did 24x7 make "terminate" attempts for these two "uncompleted" processes?
Does the job run as synchronous?
Does your EXE write to the console or it is a graphical application?
Does your EXE connect to a database or uses OLEDB or ODBC connections, other shared system resources?

Can you post fragment of your job log from 1:15 to 07:07 ?. If that is big, please email it to support@softtreetech.com.

You can turn on job tracing and profiling options to gather some additional information (Tools/Options menu, Log tab page).
If these options are checked take a look at stats.log. Check if you can see a pattern there of loosing system resources after each EXE run.

: Hi again: We had problems with our scheduled job again last night - please
: see below for the report I had to give to my manager: At 1:18am, the
: payment loader module timed out. There was no reason given, but there were
: the following lines in the log file: 11/19/2002 01:15:50 0 0 24x7
: Scheduler 24x7 Scheduler starting...
: 11/19/2002 01:15:51 0 210 Payment Loader Job started.
: 11/19/2002 01:17:26 0 197 Check CRM Job started.
: 11/19/2002 01:17:27 0 197 Check CRM Job finished.
: 11/19/2002 01:18:51 1 210 Payment Loader Timeout. Process
: 'Z:\Apps\AWBS\Script\PayLoad.bat >> z:\Apps\AWBS\output\PayLoad.log'
: has been terminated!

: Thereafter, the payment loader ran every 5 minutes as scheduled until 3:20am
: either reporting an exit code of 1 each time or not completing at all;
: i.e. at 3:00am, it was started but did not seem to complete since there is
: no accompanying "finished" message.

: At 3:30am, the Package Loader was reported to have been started also but
: there was no report of completion though the package loader seemed to
: running to completion thereafter.

: These two "uncompleted" processes would be the cause of the
: "11/19/2002 07:07:25 1 0 24x7 Scheduler 2 spawned processes are still
: running. Shutting down the scheduler now may cause unpredictable
: results." message this morning when a shutdown of the scheduler was
: attempted.

: The checks I put into the PayLoad.bat file itself did not yield any further
: information and I have since modified the checks so that time and date
: information is also recorded in a log file.