If a job creates a subdirectory, the consecutive download request creates a file rather than a directory in the resulted destination:
mkdir mydir
echo boo > mydir/demonstration
ls -lAR
Submit it, downloadclean job:
ls -lAR
.:
total 4
drwxrwx---+ 2 scop013 all 105 Oct 10 10:40 dc-dir
./dc-dir:
total 32
-rw-rw----+ 1 scop013 all 66 Oct 10 10:40 dc-dir
-rw-rw----+ 1 scop013 all 0 Oct 10 10:40 dc-dir.e47493
-rw-rw----+ 1 scop013 all 0 Oct 10 10:40 dc-dir.o47493
-rw-rw----+ 1 scop013 all 0 Oct 10 10:40 mydir <-- ERROR, not a directory
-rw-rw----+ 1 scop013 all 0 Oct 10 10:40 stderr.txt
-rw-rw----+ 1 scop013 all 235 Oct 10 10:40 stdout.txt
This is what it was on the compute server:
total 8
-rw------- 1 scop013 uoastats 0 Oct 10 10:38 stdout.txt
-rw------- 1 scop013 uoastats 0 Oct 10 10:38 stderr.txt
drwx------ 2 scop013 uoastats 4096 Oct 10 10:38 mydir <-- CORRECT
-rw-r--r-- 1 scop013 nesi 66 Oct 10 10:37 dc-dir
If a job creates a subdirectory, the consecutive download request creates a file rather than a directory in the resulted destination:
mkdir mydir echo boo > mydir/demonstration ls -lAR
Submit it, downloadclean job:
ls -lAR .: total 4 drwxrwx---+ 2 scop013 all 105 Oct 10 10:40 dc-dir
./dc-dir: total 32 -rw-rw----+ 1 scop013 all 66 Oct 10 10:40 dc-dir -rw-rw----+ 1 scop013 all 0 Oct 10 10:40 dc-dir.e47493 -rw-rw----+ 1 scop013 all 0 Oct 10 10:40 dc-dir.o47493 -rw-rw----+ 1 scop013 all 0 Oct 10 10:40 mydir <-- ERROR, not a directory -rw-rw----+ 1 scop013 all 0 Oct 10 10:40 stderr.txt -rw-rw----+ 1 scop013 all 235 Oct 10 10:40 stdout.txt
This is what it was on the compute server:
total 8 -rw------- 1 scop013 uoastats 0 Oct 10 10:38 stdout.txt -rw------- 1 scop013 uoastats 0 Oct 10 10:38 stderr.txt drwx------ 2 scop013 uoastats 4096 Oct 10 10:38 mydir <-- CORRECT -rw-r--r-- 1 scop013 nesi 66 Oct 10 10:37 dc-dir
Archiving works OK, though.