Description of problem: Only the first "ls -l" after fuse-mount a replica 2 volume using BD xlator, shows the expected values (sizes, mtime, etc). The volumes works as expected with qemu but virsh and virt-manager show lvs with a zero size. Version-Release number of selected component (if applicable): Fedora 22 GlusterFS 3.6.4 Steps to Reproduce: 1. fuse mount a glusterfs volume with some lvs exported 2. ls -lh the mount point 3. ls -lh again. Actual results: The second and subsequent ls, show the attrs of the posix brick, not the ones generated by the db xlator. Other operations show inconsistent results. Example: [root@monica mnt]# mount -t glusterfs erika:matias tt [root@monica mnt]# ls -lh tt total 514G -rw-r--r--. 1 qemu root 120G Aug 11 00:32 all_sog -rw-r--r--. 1 root root 0 Jul 28 07:48 devel -rw-r--r--. 1 qemu root 90G Aug 11 00:32 extras -rw-r--r--. 1 qemu root 300G Aug 11 00:41 mirrors -rw-r--r--. 1 root root 0 Jul 28 07:48 people -rw-r--r--. 1 qemu root 4.0G Aug 11 00:40 TamTest [root@monica mnt]# ls -lh tt total 0 -rw-r--r--. 1 qemu qemu 0 Aug 10 14:15 all_sog -rw-r--r--. 1 root root 0 Jul 28 07:48 devel -rw-r--r--. 1 qemu qemu 0 Aug 10 14:15 extras -rw-r--r--. 1 qemu qemu 0 Aug 11 00:41 mirrors -rw-r--r--. 1 root root 0 Jul 28 07:48 people -rw-r--r--. 1 qemu qemu 0 Aug 11 00:40 TamTest [root@monica mnt]# cd tt [root@monica tt]# stat mirrors File: ‘mirrors’ Size: 0 Blocks: 0 IO Block: 131072 regular empty file Device: 2eh/46d Inode: 12349220777661360721 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 107/ qemu) Gid: ( 107/ qemu) Context: system_u:object_r:fusefs_t:s0 Access: 2015-08-11 01:33:50.734322000 -0500 Modify: 2015-08-11 01:33:50.734322000 -0500 Change: 2015-08-11 01:33:50.734560428 -0500 Birth: - [root@monica tt]# stat mirrors File: ‘mirrors’ Size: 322122547200 Blocks: 629145600 IO Block: 131072 regular file Device: 2eh/46d Inode: 12349220777661360721 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 107/ qemu) Gid: ( 0/ root) Context: system_u:object_r:fusefs_t:s0 Access: 2015-08-05 12:41:47.440154000 -0500 Modify: 2015-08-11 01:34:02.702373769 -0500 Change: 2015-08-11 00:29:34.679467362 -0500 Additional info: [root@erika ~]# gluster volume info matias Volume Name: matias Type: Replicate Volume ID: 005ed81a-de13-442d-b56d-d1f4c65da71f Status: Started Xlator 1: BD Capability 1: offload_copy Capability 2: offload_snapshot Number of Bricks: 1 x 2 = 2 Transport-type: tcp Bricks: Brick1: erika:/mnt/CommonBR/bd-meta Brick1 VG: matias_uno Brick2: xiomara:/mnt/CommonBR/bd-meta Brick2 VG: matias_dos Options Reconfigured: cluster.server-quorum-type: none cluster.quorum-type: auto network.remote-dio: enable cluster.eager-lock: enable performance.stat-prefetch: off performance.io-cache: off performance.read-ahead: off performance.quick-read: off server.allow-insecure: on storage.owner-gid: 36 storage.owner-uid: 107 nfs.disable: on performance.open-behind: off Additional observation: When mounting with '-o use-readdirp=off' the problem goes away.
GlusterFS-3.6 is nearing its End-Of-Life, only important security bugs still make a chance on getting fixed. Moving this to the mainline 'version'. If this needs to get fixed in 3.7 or 3.8 this bug should get cloned.
Thank you for your report. block device translator is no longer being maintained. Hence closing this bug.