I am using TPT to load an extract. The job is running fine but it prints a lot of statements to the stdout of unix.
Job end: Mon Jul 14 10:08:43 2014
Job start: Mon Jul 14 10:08:23 2014
Update__person_mst: Rows Deleted: 0
Update__person_mst: Rows Inserted: 6
Update__person_mst: Rows Updated: 0
Update__person_mst: entering DML Phase
Update__person_mst: connecting sessions
Update__person_mst: disconnecting sessions
Update__person_mst: entering Cleanup Phase
Update__person_mst: preparing target table(s)
Update__person_mst: private log not specified
Update__person_mst: entering Acquisition Phase
Update__person_mst: entering Application Phase
Job step Load__person_mst completed successfully
Teradata Parallel Transporter Update Operator Version 14.10.00.05
Update__person_mst: End : Mon Jul 14 10:08:43 2014
Update__person_mst: Start : Mon Jul 14 10:08:27 2014
DataConnector__person_mst: TPT19221 Total files processed: 1.
Update__person_mst: Total processor time used = '1.07 Second(s)'
DataConnector__person_mst: TPT19003 NotifyMethod: 'None (default)'
Update__person_mst: The job will use its internal retryable error codes
DataConnector__person_mst: TPT19008 DataConnector Producer operator Instances: 1
Teradata Parallel Transporter DataConnector__person_mst: TPT19006 Version 14.10.00.05
Job load_table.dev.stage._person_mst..dev completed successfully
Teradata Parallel Transporter Version 14.10.00.05
Job id is load_table.dev.stage._person_mst..dev-90797, running on develop
Job log: /opt/teradata/client/14.10/tbuild/logs/load_table.
I want to either suppress these or redirect them to a log file.
Can someone suggest how this can be accomplished ? What option or code needs to be added to my tbuild script ?
I was able to redirect to a log file using a standard redirect unix operator
tbuild > xyz.log
However, just wondering if TPT has some log tracel level implementation to prevent this level of IO in case one is moving large number of tables. ?
Nopes, Jokes aside here. :)
I am loading 550+ tables and this was just a snapshot of some logs that are written by TPT.
50+ lines for every job, It seems wasteful at times and i am more keen to know the trace levels settings for bettwe management and debugging.
I don't think you can switch off the output to stdout within TPT.
The TRACELEVEL option controlling what is written to the log is set to NONE by default, setting it to a higher level might produce a really huge log (but that's not written to stdout by default).
I believe TPT supports the -o command line option to write out the "console" output to a file. The redirection works just fine as well.
That console output is just a "brief" output to show progress of the job (and we get requests from customers all of the time to add more information to the console output so that they do not have to go to the verbose logs to get job status; we try to be careful how much information we output to the console).
The TraceLevel atrtribute controls the amount of information written by the operators in their private logs (virtual logs within the .out file). TraceLevel is usually only recommended when trying to analyze a problem with a job. The default/standard output by the operators usually provide all of the necessary information for the user.
The real lengthy information is in the log (the .out file).
As Dieter indicated there is no option to disable console output.