Follow-up Comment #1, bug #87960 (project xrootd): Can you unmount everything and make sure there is no xrootdfs process running, and then mount them again? Also make sure you are not building loops. I am trying to understand your setup. In one of the other e-mail, you said atlas8 = atlas21 ------------------ 1 since on ascint1y, you mount xrootdfs /atlas21/atlas fuse rdr=root://atlas21.hep.anl.gov:1094//atlas,uid=... and you get in syslog: Oct 19 06:53:46 ascint1y XrootdFS[20560]: INFO: use the following 1 data servers : Oct 19 06:53:46 ascint1y XrootdFS[20560]: localhost.localdomain:1094 I supposed localhost.localdomain means 127.0.0.1, then I got atlas21 = ascint1y ------------------- 2 So from 1 and 2, I have ascint1y = atlas8 = atlas21, then how can the first and second df shows difference result? xrootdfs 3.6T 187G 3.4T 6% /atlas21/atlas xrootdfs 19T 163G 19T 1% /atlas21/atlas For atlas22, it seems from the syslog that it is a redirector that has three data servers: ascint2, ascint3, ascint1y. If so why bother to login to atlas22 (the redirector) and df there? _______________________________________________________ Reply to this item at: <http://savannah.cern.ch/bugs/?87960> _______________________________________________ Message sent via/by LCG Savannah http://savannah.cern.ch/