Scheduled reports not generated on Jasper Server

[bhavya_s]

Hi All,

 

I am facing an issue when I schedule 103-104 reports to run at the same time.

 

Whenever we schedule the reports to run, a database entry is made and deleted once the report is generated. When I scheduled 103-104 reports to run at the same time, not all reports were generated. This happened twice. Once I scheduled 104 reports and only 48 reports were generated. There were 56 entries in the database which were deleted once I restarted the Jasper Server. Again I scheduled 103 reports out of which only 51 reports were generated and there were 52 entries in the tables. These entries were deleted once the Jasper Server was restarted.

 

I have no idea why this is happening. I checked the log file and didn’t find any errors or warnings related to this issue. Is this because we are using evaluation version? I didn’t have this issue when I scheduled 25 reports to generate at the same time.

 

Any help is highly appreciated.

 

Thanks & Regards,

Bhavya

[swood]

Were there no error documents generated? If a job fails, it should create a document indicating an error. Could the server have run out of memory or something?

There is a configuration in the WEB-INF/js.quartz.properties file that limits the number of scheduled reports running simultaneously, so that if you had more reports than the number of scheduled reporting threads, the other scheduled report runs will wait.

 

Sherman

Jaspersoft

[bhavya_s]

Thanks for the reply Sherman.. :)

 

No Sherman.. There were no error messages either on the console or log file. If the server runs of memory, out of memory exception should appear on the console or in the log file right? I did not get any such exceptions.

 

I found the following entry in WEB-INF/js.quartz.base.properties: org.quartz.jobStore.misfireThreshold=180000. Is this what you are refering to?

 

Thanks & Regards,

Bhavya

[swood]

The thread setting is:

 

org.quartz.threadPool.threadCount=2

 

Sherman

Jaspersoft

[theodan]

Was any solution ever found for this?  It seems related to my issue as well:

 

jasperforge.org/plugins/espforum/view.php

 

It seems that something in the way JS configured Quartz makes it unable to handle anything more than a few (i.e. less than 50 or so) scheduled reports that are supposed to run at the same time.

 

-Dan

[bhavya_s]

Hi Dan,

 

Modify the following property in the js.quartz.base.properties file accordingly: org.quartz.jobStore.misfireThreshold=180000

 

The misfire threshhold is set to 3 seconds (180000). Increase it to whatever value is relevant for you. I had set it to 3600000 (one hour) and problem was resolved.

 

- Bhavya

Post Edited by bhavya_s at 06/05/2009 09:35

[theodan]

Thanks Bhavya.  I had figured that out, but it seems like a less than ideal solution.  Right now my reporting run takes around 20 minutes.  As I schedule more reports in the next few days, it will take a couple of hours.  Eventually, it will take 10+ hours, and then several days.

 

I shouldn't have to keep increasing "org.quartz.jobStore.misfireThreshold" as my runs get longer over time.  There should be a setting like "-1" to tell Quartz or JS to always handle "misfired" jobs.

 

Better yet, JasperServer's MisfireHandler should give us the option to have it actually execute misfired jobs, rather than just logging them in DEBUG mode.  Sometimes you don't care that a job is "misfired", and you just want all jobs to be executed, even if it means Quartz may fall behind (because, for example, you know your load is such that this can never happen in reality).

 

-Dan

[lucianc]

Those are good ideas, log them as bugs/feature requests so that we keep track of them.

Regards,

Lucian