I am having a problem with the job queue backing up with pending jobs because one job
never seems to finish (as far as 24x7 sees it)

I wrote a simple program that displays a message on the screen for a period of time.
The program call is: TimedMessage "Hello World" 60

Basically, it displays the message for "60" seconds. The program runs once
every 3 minutes and runs good a number of times (like 20 times or so). But after a while,
the program will run (and complete) but 24x7 shows that it is still running (in the Job Queue),
and all other jobs then just wait in the queue.

Here is my sequence of events (when it messes up):

- Program finishes, Job log indicates it's finished.
- The 24x7 Job Queue still shows that the program is "Running"
- Go to the Task Manager, and my program is not listed,
indicating it has indeed finished.
- Future jobs just pile up in the job queue and nothing gets executed.
- Try to close 24x7 and the 24x7 won't close at all. I have to use the
Task Manager to close 24x7 Scheduler.

I run the program as "detached", and still this happens.

I am running version 3.4.2.

Any ideas?

In my previous question, I mis-stated that I ran the program
in detached mode. I did not. However, I would prefer
not to.

Robert

: I run the program as "detached", and still this happens.

As a workaround you can set the job to run as asynchronous to avoid queue contention or you can put this job to a separate queue.

For testing purposes please create a batch file like
@echo start
my_program_name.exe
@echo end

Schedule this batch file instead of your program as
my_batch.bat >> my_log.log

Run it until it hangs and check if it wrote the "end" message. This way you can find out if the process hangs because of your program or because of something else.

Please keep us posted.

: In my previous question, I mis-stated that I ran the program
: in detached mode. I did not. However, I would prefer
: not to.

: Robert

Hi,

I did everything you suggested and the same problem
is still happening. I'm totally confused since the
program is running in Async mode.

I can give you the Schedule.log file, the program I'm running,
the .bat file I ran (to run my program), the output
file that the test .bat file produced, screen prints
at the time the scheduler hangs, job properties, etc.

There must be something quite simple I'm overlooking.

Please advise. I cannot continue with 24x7 unless I
know what's happening.

Robert

: As a workaround you can set the job to run as asynchronous to avoid queue
: contention or you can put this job to a separate queue.

: For testing purposes please create a batch file like
: @echo start
: my_program_name.exe
: @echo end

: Schedule this batch file instead of your program as
: my_batch.bat >> my_log.log

: Run it until it hangs and check if it wrote the "end" message. This
: way you can find out if the process hangs because of your program or
: because of something else.

: Please keep us posted.

An async job can clog the queue if it has "on finish" or "on error" notification action and that action hangs. Please let me know what you have for this job.

Another remote possibility is that the job keeps loosing system resources in particular number of system handles or threads which is limited to 64K on any Windows system. Once the job (actually all running jobs together) hit that limit the scheduler runs out of system resources and hangs.
You can use the Task Manager to monitor handles and threads. You may need to customize Tasks Manager view and make these columns visible (while on the Processes tab, click View/Select Columns menu). If these numbers keep incrementing all the time you sure have some kind of resource leakage issue.

: Hi,

: I did everything you suggested and the same problem
: is still happening. I'm totally confused since the
: program is running in Async mode.

: I can give you the Schedule.log file, the program I'm running,
: the .bat file I ran (to run my program), the output
: file that the test .bat file produced, screen prints
: at the time the scheduler hangs, job properties, etc.

: There must be something quite simple I'm overlooking.

: Please advise. I cannot continue with 24x7 unless I
: know what's happening.

: Robert

Hello again,

I've been away from my 24x7 tasks for a few days, but
I'm still having the same problem with my test app. I
monitored the resources on my machine (Win 2000 Server)
and my handles and threads always returned to the
same amount after my job had completed. This is a
test server with really nothing running on it besides
24x7 and has 1/2 gig memory.

The only thing my job does is launch the test
app (which displays "Hello World" for 10 seconds) and
then quits. It is launched once every minute and
the job has no notifications at all (as you suggested may
clog up the queue). Actually, this is the only job 24x7
has active. In short, this is the most simple
example I could come up with and I can't seem
to get the scheduler to cooperate. IMO, this is a
definate show stopper!

Any more suggestions?

: An async job can clog the queue if it has "on finish" or "on
: error" notification action and that action hangs. Please let me know
: what you have for this job.

: Another remote possibility is that the job keeps loosing system resources in
: particular number of system handles or threads which is limited to 64K on
: any Windows system. Once the job (actually all running jobs together) hit
: that limit the scheduler runs out of system resources and hangs.
: You can use the Task Manager to monitor handles and threads. You may need to
: customize Tasks Manager view and make these columns visible (while on the
: Processes tab, click View/Select Columns menu). If these numbers keep
: incrementing all the time you sure have some kind of resource leakage
: issue.

I still think that the problem is in your program.
Email your "Hello World" test app to support@softtreetech.com so we can test it and confirm that your application does really exist after the specified time.

While we are inveestigating this problem set the job to run synchronous and set let's say 15 seconds timeout so 24x7 can kill your process if it does not finish in time.

: Hello again,

: I've been away from my 24x7 tasks for a few days, but
: I'm still having the same problem with my test app. I
: monitored the resources on my machine (Win 2000 Server)
: and my handles and threads always returned to the
: same amount after my job had completed. This is a
: test server with really nothing running on it besides
: 24x7 and has 1/2 gig memory.

: The only thing my job does is launch the test
: app (which displays "Hello World" for 10 seconds) and
: then quits. It is launched once every minute and
: the job has no notifications at all (as you suggested may
: clog up the queue). Actually, this is the only job 24x7
: has active. In short, this is the most simple
: example I could come up with and I can't seem
: to get the scheduler to cooperate. IMO, this is a
: definate show stopper!

: Any more suggestions?

Hi,

I think I have isolated my problem (hopefully).

FWIW, I took my program and put it on two other
machines running 24x7. They ran without incident.
Just by chance, I went into the 24x7 options and
saw that the software was configured to run as a
service. I know that I had manually stopped the
service several days before so I could work with the
interface without having to use it in remote control
mode. However, the configuration was still that
24x7 was to run as a service (even though I was
running it standalone while the service was stopped).

In short, what I did was to configure 24x7 not to run
as a service and the program worked without incident. Then,
out of curiosity, to duplicate my previous errors, I configured the
24x7 to run as a service again, rebooted, then stopped
it manually. Then I ran 24x7 in standalone mode, and
sure enough my most simple program started to misbehave.

Sorry for the long post, but I guess I fell into an
unlikely trap. I thought if you stopped the service,
you could just go ahead and use the software standalone.

Did this make any sense?

Robert

: I still think that the problem is in your program.
: Email your "Hello World" test app to support@softtreetech.com so we
: can test it and confirm that your application does really exist after the
: specified time.

: While we are inveestigating this problem set the job to run synchronous and
: set let's say 15 seconds timeout so 24x7 can kill your process if it does
: not finish in time.

Stopping the service does not stop the 24x7 from behaiving as a "graphical interface only." Anyway, I am grald you were able to solve the issue.

: Hi,

: I think I have isolated my problem (hopefully).

: FWIW, I took my program and put it on two other
: machines running 24x7. They ran without incident.
: Just by chance, I went into the 24x7 options and
: saw that the software was configured to run as a
: service. I know that I had manually stopped the
: service several days before so I could work with the
: interface without having to use it in remote control
: mode. However, the configuration was still that
: 24x7 was to run as a service (even though I was
: running it standalone while the service was stopped).

: In short, what I did was to configure 24x7 not to run
: as a service and the program worked without incident. Then,
: out of curiosity, to duplicate my previous errors, I configured the
: 24x7 to run as a service again, rebooted, then stopped
: it manually. Then I ran 24x7 in standalone mode, and
: sure enough my most simple program started to misbehave.

: Sorry for the long post, but I guess I fell into an
: unlikely trap. I thought if you stopped the service,
: you could just go ahead and use the software standalone.

: Did this make any sense?

: Robert