DSXchange

I've met two problems when using oracle load (parallel=true,direct=true,remote server,dataset to oracle or transformer to oracle)

1, there are a lot error mixed warning messages in logs, it overflow the rows count that can show in director. how can easy to get all the logs from the starting of job?

2, error message repeats:
....
---The call to sqlldr faild; the return code=35072;()
---Unable to access the log file.
---The runLocally() of the operator failed.
---Input 0 consumed 1 records.
---Operator terminated abnormally: runLocally did not return APT_StatusOK
---/usr/lib/pa20_64/dld.sl:'opt/datastage/Ascential/DataStage/DSEngine/java/jre/lib/PA_RISC2.0/hostpot/...
---sh:3254 killed
......
has some met this error before? any help is appreciated

Lynn

The first thing to get is the sqlloader log file from the scratch disk area. This should have more information in it if it pertains to something that happened during the load. Post that information here.

Sean

You should also have the permission to the directory whereever your log file exists.
This needs to be set in the APT file for that project.

The Log view in Director is filtered to the most recent 100 rows by default, to limit the amount of data that must be sent back to the client every time the Director refreshes. To change this invoke the Filter (Ctrl-T, or right-click in the background area). The row count filter is at the bottom of the dialog. Refresh interval can be controlled from Tools > Options.

i can't fould load log at scratch disk, looks like it didn't generator load log file.

the warning message begin with:

TB_TMP_USER_ORDER_INFO,1: /usr/lib/pa20_64/dld.sl: '/opt/datastage/Ascential/DataStage/DSEngine/java/jre/lib/PA_RISC2.0/hotspot/libjvm.sl' is not a valid load module: Bad magic number

i think this is the fatal reason.