There are 50 Autosys jobs that runs in sequence.
Now, if a job in the sequence is failed. Then I am looking for a way to manually start the failed job.
It could be easily done by calling sendevent command. But, in Production environment, we have to raise WR for that.
So, how to restart failed job manually without sendevent command?
Possible solution is to make each job dependent on a File Watcher job. But, this way, we have to create File Watcher for each job. Is there any better approach.
The only way you can manually start a job is by sendevent. But if you want to skip that, then you can do one thing. But for this, you need to have access to either one of this.
If you have this, then you can create a job, which runs every 5or 10 minutes. This job will kick off a shell script, say startJobs.sh.
In the shell script, you read a file, say jobsToStart.txt, which contains list of jobs to start. Then you use a sendevent in this script with the job name from the file.
Now once this script is deployed to production, you just need to put a job name in the file jobsToStart.txt and the script will start the jobs when the it runs.
Another way would be similar, but instead of putting data in a file, you can put data in a database table. The shell script will read that table to find the job name.
Personally, I would suggest to put n_retrys=1 to automatically restart the job and if it fails again, have the support team do it. There is a reason for access restriction and if you feel that you need to be able to do it on your own, then you have to present your case to business to give you access.
