When encountering an issue during execution, please do due diligence to check whether the source of the error is in the user script/tool. If you suspect the source of the error comes from EnTK or tools below please open a ticket in the github repo and follow these steps:
Run your script again with RADICAL_VERBOSE=DEBUG and
RADICAL_ENTK_VERBOSE=DEBUG. Once these environment variables are set, a lot
of messages will be displayed as they are written to the standard error
stream. Please redirect these messages to a single file.
Example:
RADICAL_VERBOSE=DEBUG RADICAL_ENTK_VERBOSE=DEBUG python example.py &> verbose.log
Attach verbose.log to the ticket.
NOTE: If you suspect sections of this file to be pointing to the error you may consider mentioning that in the ticket via inline comments.
Example:
```
These lines of the log might be
talking about the error
```
When running a RP script or a tool that uses RP, multiple logs are created by the components of RP. A set of these logs are created in the current working directory on the client machine (where your script lies) and a set of logs are created on the remote machine (HPC) in a specific location. You can bring all the logs to the client by running the following cmd (on the client):
radicalpilot-fetch-logfiles <session id>
In order to determine the session id, you can look for a folder that is created on the client in
current working directory. It should have the format rp.session.*. You can find the latest
folder by doing ls -ltr (last is recent).
All the logfiles are brought to this rp.session.* folder. Please zip this folder and attach to github ticket.
Sometimes the above two steps may not be enough, the developers might require to take a look at the entire sandbox. You may either downlod the entire sandbox from the remote machine (using scp/gsiscp), zip the folder, and attach it to the ticket or provide access to the sandbox on the remote machine.