Mload failure

Tools & Utilities

Mload failure


I'm facing an issue with mloading of a table via Informatica - I get a "Broken Pipe" error which I believe is a very common error. I tried the usual steps of releasing the mload lock and deleting the work tables but the issue is still unresolved. I'm working in a shared and controlled environment and need to reach out to the DBA or other teams to delete anything for us so I cannot try out anything much myself. Can someone help me identify what could be the source of this issue and where I should be looking? I've pasted the loader log below.

Thanks in advance for your help.

Teradata Parallel Transporter Version
Execution Plan generation started.
Execution Plan generation successfully completed.
Job id is r2_agg_rx_pres_prod_grp1_out-2933, running on
Found CheckPoint file: /opt/teradata/client/tbuild/
This is a restart job; it restarts at step MAIN_STEP.
Teradata Parallel Transporter DataConnector Version
PRODUCER_OPERATOR: Instance 1 directing private log report to 'PL_R2_AGG_RX_PRES_PROD_GRP_out-1'.
Teradata Parallel Transporter Update Operator Version
CONSUMER_OPERATOR: private log specified: CL_R2_AGG_RX_PRES_PROD_GRP_out
PRODUCER_OPERATOR: Instance 1 restarting.
PRODUCER_OPERATOR: DataConnector Producer operator Instances: 1
CONSUMER_OPERATOR: connecting sessions
CONSUMER_OPERATOR: preparing target table(s)
CONSUMER_OPERATOR: entering Acquisition Phase
PRODUCER_OPERATOR: Operator instance 1 processing file '/fproc/infa_repos_stst/TgtFiles/r2_agg_rx_pres_prod_grp1.out'.
PRODUCER_OPERATOR: pmRepos failed. General failure (34): 'pmUnxDskSetPos: 'seek' error (Illegal seek)'
PRODUCER_OPERATOR: Fatal error repositioning data.
Operator( instance(1): RESTART method failed with status = Fatal Error
PXTB_Restart: Operator restart error, status = Multi Phase Error
Task(SELECT_2[0001]): restart completed, status = Multi Phase Error
CONSUMER_OPERATOR: disconnecting sessions
PRODUCER_OPERATOR: Total files processed: 0.
Job step MAIN_STEP terminated (status 12)
Job r2_agg_rx_pres_prod_grp1_out terminated (status 12)
Teradata Employee

Re: Mload failure

The post subject line says "Mload failure" but this is really TPT, not MultiLoad.

If this job will always fail when re-running it, then a few things can be done:

1. go into $TWB_ROOT/logs and look for a file called "r2_agg_rx_pres_prod_grp1_out-2933.out"
2. send that to me.
3. or you can run the following:

tlogview -j r2_agg_rx_pres_prod_grp1_out-2933 -f "*" -g

this will process that .out file and provide all of the information from all of the operators in that log

The thing is, no diagnostics were turned on, so we may not see much relevant info.

Next, if it is possible to modify the script that was being run, then you want to set a TraceLevel attribute in the
DataConnector operator to 'all':

VARCHAR TraceLevel = 'all'

This will enable runtime diagnostics in the DataConnector operator (file reader) which may tell us why the seek is

Also, are you sure that "r2_agg_rx_pres_prod_grp1.out" is the correct name of the data file?
-- SteveF
Teradata Employee

Re: Mload failure

BTW, if you release the locks, and you delete the work tables, all previous rows sent to Teradata
are gone. This means you will have to start the job from the very beginning, meaning you need to
also delete the error tables, restart log table, and the TPT checkpoint files.

-- SteveF

Re: Mload failure


Thanks for your inputs, turns out the DBA had not deleted some of the error tables which led to this failure. Sorry for the incorrect reference to Mload, our systems were recently migrated from Oracle to Teradata and we have just started learning this new technology. I'll definitely refer to these instructions whenever I come across such a situation in future.

On a side note - we do not have access to edit the TPT script, and the out filename is correct.