1. "If the job paused during the application phase, and the work and error tables have not been dropped and restart log table is dropped, then restart the job" (I changed it as per my understanding -- Mload reference 14.10 pg no. 42)
2. "If the Teradata MultiLoad job script has been lost, use the following abbreviated script
to restart the job:
.BEGIN MLOAD . . . ;
The BEGIN MLOAD command must identify all of the target tables, work tables, and
error tables of the original job, and they must all be intact and not corrupted. When
submitting the abbreviated script, Teradata MultiLoad uses the checkpoint markers in the
work tables to restart the task at the point it stopped in the application phase."
So, can we restart a Multiload job that paused/failed in application phase even without Restart log table? So, if we restart the job, it tries to read the restart log table which is not present and hence will try to read the work tables? Can the Mload job be restarted without restart log table even in Acquisition phase? Are these check point markers produce equilvalent information that should be in restart log table? Thank you for your time on this.
It is possible that the job will not run because the job will be out of sync.
That is how it is supposed to work.
MultiLoad will get an indication from Teradata that the job is a restarted job.
And yet there is no log table.
Therefore there is a conflict and the job should terminate with a fatal error.
In the 2nd point(in case of loss of script), the Teradata MultiLoad uses the checkpoint markers in the work tables to restart the task at the point it stopped in the application phase.
So, the restarted job will read the restart log table and should start with where it left off. Now as stated in above read check point markers in work tables, so how different is this(check point markers in WT) to that of information in restart log table? I hope that the restart log table is updated on every block written to target table and checkpoint markers in WT is written for each row written to target table. Please correct me if I am wrong in this!
I see that you are posting several questions on utility/DBS restarts.
Please refer to the utility documentation.
The restart log table will keep track of what phase the job was in. During the acquisition phase it will also keep track of the checkpoints, if there are any. The application phase is a DBS-only action. There is nothing in the restart log table to keep track of, other than the fact that the job was in the application phase.
If there is a DBS restart during the application phase, it is the DBS's responsibility to determine where to resume. The client utility has no control over that.
The client utility and the restart log table must always be in sync with each other.