Putting all messages in one log file
Posted: Mon Nov 22, 2010 12:38 pm
I've been looking for good ways to avoid needing to log on to Director for fatal error messages. Ray D.'s recent post about Director Lite is a step up from my current situation, but not likely to be used any time soon. Before I describe the details, I also need to point out that we do not put design components into production, executables only, so all of the nice metadata functions described in the manuals give me very little to work with. Sorry about the bad news up front. 
Basic details, hoping I don't leave out something important: Batch scheduling is done exclusively in Control-M. CM command line executes a generic script with the project name, sequence name, etc. as input parameters. The script uses standard Unix logging functions with log files put in the usual path.
I need a reliable method of identifying every job that runs under each sequence. I could create a text parameter file that lists every job name and use that to retrieve from the logs, but that also means an extra item to deploy and to manually maintain if sequences change. I have a small app that uses this. I'm looking for something I can use that avoids that sort of solution. If you are going to suggest using invocation id, please go into some detail about it. It seems promising but I've not gained a good understanding of it yet.
Finally, I'm woefully ignorant of the inner workings and details of the DS Engine. I haven't had much opportunity to get into it. If there is a document floating around that can get me started with that, I'd rather read it first before bugging you all with this. I am also an experienced developer, so I'm not looking for any help in Unix or DataStage coding basics. Command and syntax with brief explanations will do nicely, with my gratitude.
Basic details, hoping I don't leave out something important: Batch scheduling is done exclusively in Control-M. CM command line executes a generic script with the project name, sequence name, etc. as input parameters. The script uses standard Unix logging functions with log files put in the usual path.
I need a reliable method of identifying every job that runs under each sequence. I could create a text parameter file that lists every job name and use that to retrieve from the logs, but that also means an extra item to deploy and to manually maintain if sequences change. I have a small app that uses this. I'm looking for something I can use that avoids that sort of solution. If you are going to suggest using invocation id, please go into some detail about it. It seems promising but I've not gained a good understanding of it yet.
Finally, I'm woefully ignorant of the inner workings and details of the DS Engine. I haven't had much opportunity to get into it. If there is a document floating around that can get me started with that, I'd rather read it first before bugging you all with this. I am also an experienced developer, so I'm not looking for any help in Unix or DataStage coding basics. Command and syntax with brief explanations will do nicely, with my gratitude.
