Some potential improvements and new features for CandyPond, in particular the arccandypond tool. To be discussed with @andrii
- Set PYTHONPATH in the RTE so that jobs can use the python API of arccandypond
- Enable a timeout to
getso that jobs do not wait too long in case A-REX is too busy with data staging. The job may want to fall back to direct copy in this case.
CacheLinksupports multiple transfers in one "job", so it makes sense to send the requests in bulk rather than calling sequentially. Either this could be reading lines of "url file" from a file or specifying extra arguments to arccandypond like
arccandypond get url1 file1 url2 file2 ...
- Is arccandypond python3 compatible?