I recently implemented two new features into gridtk:
1. The first feature (which I actually implemented second) allows to add time stamps into the log files, when the job is started and when it finished.
The time stamp including command line is written into the error log file (the standard way of logging with python).
It relies on the -vv option during submission of jobs (in grid mode) or during running the local scheduler (in local mode).
2. I implemented a feature that allows to run the script in a given directory, instead of in the current directory.
You can use the --exec-dir option during submission of the jobs.
Theoretically, this feature should be applicable to both modes (grid or local).
Notes:
A. Unfortunately, I was not able to test the new functionality in the grid, as we don't have an SGE here.
B. As I had to modify the API of the Job model inside gridtk, the new gridtk won't work with the old sql databases, and vice versa.
I haven't yet pushed the new version of gridtk, as I wanted first to have your opinion about the changes (particularly the API change) and second I need someone to test the new functionality in the grid. Any volunteers?