GitLab

Gets the birth/death events instead of sampling the database. The "NOTHING" curve is not needed any more

Only gnuplot has to know about the filtered analysis, the CSV file should still contain all the data

schema_change: detect and register RELOCATED events that used to mess up things on LSF 9.0 ("job rescheduled" in LSF parlance)

be more careful with fetch_overdue_workers (Use 5sec threshold to avoid checking recently active Workers. Do not use it at all when performing -all_dead.)

bugfix:  last_check_in should only be updated by register_worker_death if the Worker is burying itself

cosmetic: added (commented out) warning messages for every external system() call that LSF module runs - simplifies debugging a lot

ResourceDescription expanded to include both submission_cmd_args and worker_cmd_args. Both args can be specified in a PipeConfig file.

(1) do not change SEMAPHORED jobs to READY and (2) support more flexibility in choosing which statuses to reset

schema change: turned all VARCHAR(<255) into VARCHAR(255) -- should improve experience with long host namest (thanks, MichaelP!)

param_required() now automatically sets transient_error(0) before dying, to avoid unnecessary retries (thanks, Matthieu!)