On 2/4/20 5:22 PM, Adrian Sevcenco wrote: > On 2/4/20 4:23 PM, Michal Kamil Simon wrote: >> Hi Adiran, > Hi! Hi! So, i closed the ticket as, playing with metalink file the xrdcp was well behaved and it was doing the fallback to the next source if the first url was broken At this moment i begun to doubt my initial observation (when using the python bindings) that the fallback was not happening, but i will try to gather some more debug logs when/if it happens again. I take the change to hijack the thread and as about the parametes: checksummode = 'none', checksumtype = '', from the copyprocess.py i tried to use them as target, md5 and in the log i have : [2020-02-05 01:14:52.361230 +0200][Debug ][Utility ] Attempting checksum calculation, mode: target. but nothing seems to work as the file is downloaded and kept nevertheless (i modified by hand the md5 sum from metafile to simulate the failure). How should i use these parameters? I would expect that the md5 from metafile (defined by checksumtype) would be compared with the fresh computed md5 of the downloaded file and if not equal the file should have been deleted... Should i open an issue? Thank you!! Adrian ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the XROOTD-L list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1