Description of problem: gluster volume log filename command is used to set the log file for corresponding volume/brick. When used, it creates a directory with the same name instead of a life. Version-Release number of selected component (if applicable): 3.2.5 How reproducible: gluster volume log filename volume-name brick-name /var/new-log.log Steps to Reproduce: 1. gluster volume log filename amitest ec2-174-129-105-24.compute- 1.amazonaws.com:/mnt/x4 /var/log/glusterfs/bricks/mnt-x4-testing.log Actual results: [root@ip-10-116-242-85 ~]# gluster volume log filename amitest ec2-174-129-105-24.compute-1.amazonaws.com:/mnt/x4 /var/log/glusterfs/bricks/mnt-x4-testing.log log filename : successful [root@ip-10-116-242-85 ~]# ls /var/log/glusterfs/bricks/mnt-x4-testing.log/ [root@ip-10-116-242-85 ~]# tail /var/log/glusterfs/bricks/mnt-x4-testing.log tail: error reading `/var/log/glusterfs/bricks/mnt-x4-testing.log': Is a directory Expected results: /var/log/glusterfs/bricks/mnt-x4-testing.log should be the new logfile. Additional info: help text for the command: volume log filename <VOLNAME> [BRICK] <PATH> - set the log file for corresponding volume/brick
Currently, this can be documented as a known issue. Because, the behavior of 'log filename' is to set the directory for the brick, and file name is always the brick's path... In future (ie, 3.3.x onwards log-filename will not be available in CLI command set).
glusterfs-3.3.0 doesn't have 'gluster log filename'. Hence moving it as CLOSED UPSTREAM