sempahores not being posted

Hi Everyone,
we have encountered this issue with 24*7 where jobs would not post sempahore at all and other jobs waiting for the semaphores would not launch. Time based jobs seems to work fine. But once or twice a week, we have this issue of jobs not posting semaphores. Has anyone encountered this issue before? We are running 24*7 on a windows machine. I would appreciate any replies.

Thanks,
Ro

Most likely you're dealing with Windows UAT and file virtualization issues and not running the scheduler as a Windows service (you are running I in Desktop mode).
If that's the case, please restart the scheduler in Run as Administrator mode or as a Windows service.

Restarting the machine or 24*7 services always fix this issue but we are trying to figure out a permanent solution. I would like to know what is causing this to happen and do a root cause analysis.

If there are no errors and exceptions reported in debug.log and main log files, the scheduler isn't aware of the issue, the operation system or middleware doesn't report them.

Please consider enabling the tracing option in the scheduler settings. Reproduce the issue and review debug.log for related exception messages if any. Please share with us what you will find out.

I suggest creating a heartbeat job using built in JavaScript engine that would try to delete then create some dummy text file, then check if file exists, and if not, send an email alert. The job needs to be run non asynchronous and non detached. You can set it to run every 5 minutes or so. Something like the following concept script.

If File.exists(...) {File.delete(...);}
File.saveAlll('dummy text', ...);
If !File.exists(...) {Email.send(...,...);}

This way you can get an automatic notification when the issue occurs again.

Hi,

Debug.log doesn't have much information and the time 24*7 didn't post the semaphore there was no log details in debug file. Where can I enable the tracing option in the scheduler settings? I would appreciate if you can share that information.
We have an alert that sends us notification when the semaphores stopped posting but it is happening too often and we have to run a lot of jobs after we reboot the services.
Also, we have 24x7mp_setup_5.3.457 version of 24*7 and not the latest one. Should we update Java to latest as well?
A colleague of sent an email few months ago and here was the response..

"Would you be able to setup a performance monitor for open file handles and check where it goes when the file operations start getting impacted?"

Perf mon has not been very helpful either.

Thanks,
Rohit

If you have access to UI:
Stop the service.
Start desktop version.
Click Tools/Options menu.
Click Log tab.
Select Trace enabled checkbox.
Click Ok.
Close the UI and restart the service.

If you use web console:
Login to server.
Click Tools menu.
Select Options link.
Locate Trace Enabled option.
Change it and click Ok.

Please let us know what you see in the log when it happens again.

can you also tell us what verison of Java we should be using with 5.3.457 verison of 24*7? Is this can cause issue with semaphores files being not posted?

You can use either version 7 or 8.

I hope you would be able to catch relevant exceptions in the debug.log which we can use to figure out the root cause of the issue.