Hide Forgot
few issues with scale-n-defrag script: *files from root dir (/mnt) files are not distributed. *if scale-n-defrag fails cd command (ex cd /mnt/ ) with nfs-stale error message. then defrag script executed on current directory - which is not glusterfs *log files are created for each file instead of directory,like ------------------------------- Completed directory /data/test2/files/f1399 Defragmenting directory /data/test2/files/f14 (/data/defrag-store-f14.log) Completed directory /data/test2/files/f14 Defragmenting directory /data/test2/files/f140 (/data/defrag-store-f140.log)
This bug is open for 'DP' (Documentation Pending). As of now, there is no revalidate sent on 'root' inode in 3.0.x, hence the issue of ESTALE when the scale-n-defrag is run on the mountpoint directly.
We have marked our 3.0.x documentation as out of date, hence no more update is happening here. This issue is not seen in 3.1.x releases, hence closing the bug. http://www.gluster.com/community/documentation/index.php/Gluster_3.0.x_Documentation http://www.gluster.com/community/documentation/index.php/GlusterFS_User_Guide
As updated in the previous comment, we are not planing to update docs for 3.0.x release, and hence DP -> DNR..