Most of the times, DataStage jobs experience missing or corrupted data. Today, in this article, we are going to discuss ‘Steps to Recover DataStage Jobs”. In this DataStage tutorial, we are going to discuss the smooth recovery ways to fix the corrupted jobs. Follow the below steps.
The Director list of jobs queries the server-side DataStage project DS_JOBS while the Designer tree view queries the xmeta repository. If the job appears in Designer but not in Director, it means the job is in the xmeta repository, but not in the UV repository or they have become corrupted.
Steps to Recover DataStage Jobs Follows
Check whether the jobs exist in the UV for the project:
- Log on to DS Administrator
- Projects tab, select the project
- Click on Command
- Type the command: LIST DS_JOBS
If the list of jobs displays, perform the following:
If the jobs exist but do not display in the clients, it usually indicates the indices are corrupted in the project. So, you need to re-index the project, as follows:
- Make sure there are no connections to DataStage (all users will have to remain off the system until this process is complete or it will not work).
- Go into the Administrator as root or dsadm
- Select the project – select command
- At the command line, execute DS.REINDEX ALL (in upper case).
- Click “Next” and “respond” each time one is active. Then, click on “close” button when active.
If the list of jobs do not display, most likely the project has become corrupted. The best way to resolve this problem is to:
- Create a new project
- Export the jobs from the problematic project
- Import them back in the new project
- Compile the new job
If you can’t afford to restore the DB backup anywhere else, you can restore at least the Depot content (NSH Scripts, Packages, etc.) from File Server. Use them to recreate your jobs/content.
If you want to identify the user deleted, go through the DB Tables and records for the same.
Overall, What Steps Should Be Taken To Recover DataStage Jobs?
In order to recover presentation of DataStage jobs, we have to first create the baselines. Secondly, we should not use only one flow for presentation testing. Thirdly, we should work in growth. Then, we should appraise data skews. Then we should separate and solve the problems, one by one. After that, we should allocate the file systems to take away bottlenecks, if any. Also, we should not embrace RDBMS in the start of the testing phase. Last but not the least, we should understand and evaluate the available tuning knobs.