We learned some more about this bug. It relates to issue #1366. There are three ingredients for this problem: `gfal-copy` trying to write files, the `root:` protocol, and HDFS as the file system. The sequence of events is as follows. 1. A request is made by `gfal-copy` to write a file using the `root:` protocol. 2. At the data node, this request fails due to #1366. The file name is added to the `posc.log`. 3. The sysadmin re-starts `xrootd` on the data node (for maintenance or other reasons). 4. During start-up, `xrootd` goes through the `posc.log` to clean up the files. It tries to delete the files from step 2, but once again it fails because of bug #1366. 5. Because `xrootd` can't clean up the files in the `posc.log`, it enters a non-responsive state and does nothing. -- Reply to this email directly or view it on GitHub: https://github.com/xrootd/xrootd/issues/1721#issuecomment-1184845715 You are receiving this because you are subscribed to this thread. Message ID: <[log in to unmask]> ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the XROOTD-DEV list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1