Interface 1: new server log message: Failed to load job script for job jobid from PBS datastore
Visibility: PBS Public
Change Control: Stable
Details: Server log, error level
When sending the job from server to mom, the server loads the job script from the database. If the loading fails due to database issues, then the above error message is shown. Additional debug information is usually available in the database logs.
Interface 2: new server log message: Out of memory loading script for jobid from PBS datastore
Visibility: PBS Public
Change Control: Stable
Details: Server log, error level
When sending the job from server to mom, the server loads the job script from the database. If the server runs out of memory while allocating space for the job script (especially if the script is large), then this error message is printed in the server logs. Job submission fails.
Interface 3: new server debug(3) message: <jobid>; big job files, sending via subprocess
Visibility: PBS Private
Change Control: Stable
Details: Server log, debug3 level
When sending the job from server to mom, the server loads the job script from the database. If the job script is larger than 2 MB in size, then the server does not send the job via TPP, rather creates a subprocess and sends the job via a direct new TCP connection from the server to the mom. This debug log message is printed in that case.
The message is also printed if job spool files (.OU, .ER and .CK) are larger than 2 MB in size in case of a job rerun.
Interface 4: new server debug(3) message: Job has no script loaded!! Can't write temp job script
Visibility: PBS Private
Change Control: Stable
Details: Server log, debug3 level
This is really to catch a programming error, i.e., if the svr_create_tmp_jobscript() is called without loading the jobscript first.