FAQ:How do I use gdb with nuwa?
| Offline Documentation: [Offline Category] [FAQ] [Howto] [Reference] [Manual] |
| Offline Documentation |
| This article is part of the Offline Documentation |
| Other pages... |
|
Offline Category |
| See also... |
Running nuwa.py under GDB
There are several ways to use gdb to debug problems with the NuWa software. They all boil down to running nuwa.py under GDB
Here, it is assumed the problematic job is simply:
nuwa.py -n 10 input.root
Start the job with GDB on the shell command line
If you know the job is going to crash, this is the easiest way to re-run it under gdb. You tell gdb the command line with "--args" flag. And, because it needs the full path to both the python executable and the nuwa.py script you can use your shell to expand them:
shell> gdb --args $(which python) $(which nuwa.py) -n 10 input.root
Attach to an already running job
This is useful if the job is already running and you want to connect gdb to it. This will be useful to debug infinite loops.
Start the job and record the process ID:
shell> nuwa.py -n 10 input.root > log 2>&1 & [1] <PID>
Or, if you don't have the PID, get it:
shell> ps auxww|grep nuwa.py (check PID)
Now connect to the PID:
shell> gdb $(which python) <PID>
Related links and info
- See instructions in Trac ticket #347
- General gdb info
- FAQ:How much memory is nuwa using?
- Dealing_With_Memory_Leaks
- Special info for PDSF
| Offline Software Documentation: [Offline Categories] [FAQ] [Offline Faq Category] |