I’m trying to understand why a batch job which will take only 3-5 seconds to run when being invoked through the DS administration panel takes 50-75 seconds to run when being called through a batch file generated with the “Export Execution Command” option on the DS admin panel.
This is running on Windows Server 2008, DS version 12.2.2.0
I need to execute the same batch job in series 5 times for 5 different server configurations, however i need the 3-5 second run time, not the 50-75 second run time as i need to run this job as frequently as possible.
Also to add, the individual jobs in the DS admin log are still only showing 3-5 seconds. It seems there is some overhead or something when the batch file is running which is causing the additional time??
The only thing I can think of is that the AL_RWJobLauncher.exe program is causing problems. Would it be possible to instead call the job through web services? Or maybe code the call using al_engine.exe instead of AL_RWJobLauncher.exe? Neither of these options is simple.
You can edit the Export Execution Command file (.bat/.txt) or create your own based on an existing file to call a job multiple times in series but with different parameters. Or, you could create a controller batch job that is in charge of calling other batch jobs using the EXEC() function that runs the .bat files.
Most of my clients use an enterprise scheduling tool. Those that do it all through DS are usually pulling their hair out. It is possible to do it all in DS, but it isn’t a simple process. A lot of thought has to go into how it is architected.
But even using the export execution command and using the batch file for Windows server still seems to have problems like i’m having here…i could use another scheduling tool to setup the sequence of events, but there seems to be extra overhead with these batch files…like you said, probably around the EXE being launched to start the job?
Correct. Even a third party scheduling tool would have the same problem since they just call the .bat file. However, I have not observed this issue in any of my clients implementations. At least not that I can recall.
If you search the forum you’ll find a few postings on how to use AL_ENGINE to start your jobs instead of the executable used in the .bat file. I don’t know that it will be any faster but I think AL_ENGINE will skip a step.
There is also the possiblity that the way AL_RWJobLauncher.exe is communicating with your job server is the problem. Maybe you have a network issue. Instead of using a server name in the .bat and .txt file maybe change it to an IP address or even LocalHost. This is wild speculation on my part.