diff options
Diffstat (limited to 'doc/release-notes')
| -rw-r--r-- | doc/release-notes/3.10.0.md | 26 |
1 files changed, 20 insertions, 6 deletions
diff --git a/doc/release-notes/3.10.0.md b/doc/release-notes/3.10.0.md index e3eb1ab9e25..f04c8de4f5e 100644 --- a/doc/release-notes/3.10.0.md +++ b/doc/release-notes/3.10.0.md @@ -153,14 +153,28 @@ manually delete the directory from the mount point. ### Implemented parallel readdirp with distribute xlator *Notes for users:* -Note is WIP! -Improves directory enumeration performance in large clusters, by performing -readdir ahead in parallel. +Currently the directory listing gets slower as the number of bricks/nodes +increases in a volume, though the file/directory numbers remain unchanged. +With this feature, the performance of directory listing is made mostly +independent of the number of nodes/bricks in the volume. Thus scale doesn't +exponentially reduce the directory listing performance. (On a 2, 5, 10, 25 brick +setup we saw ~5, 100, 400, 450% improvement consecutively) + +To enable this feature: +```bash +# gluster volume set <VOLNAME> performance.readdir-ahead on +# gluster volume set <VOLNAME> performance.parallel-readdir on +``` + +To disable this feature: +```bash +# gluster volume set <VOLNAME> performance.parallel-readdir off +``` -To test the benefits of this this feature set the following 2 volume options, +If there are more than 50 bricks in the volume it is good to increase the cache +size to be more than 10Mb (default value): ```bash -# gluster volume set <volname> performance.readdir-ahead on -# gluster volume set <volname> performance.parallel-readdir on +# gluster volume set <VOLNAME> performance.rda-cache-limit <CACHE SIZE> ``` *Limitations:* |