Description of problem: I create a distribute volume and mount to a dictionary. Then I put 25,000 files to the dictionary. After finished, I write a program to list the dictionary. I found the program spent 20s. It's unbelievable. And I copy the dictionary to root dictionary. And running the program again. The time display just less than 1s. So I think there is some problems in glusterFS. Then I do some more test for glusterFS. I found that the spent time is normal when the dictionary contains 20,000 files, but when the number is more than 20,000, it's easy to show bad performance. Finally, I found the reason of bad performance is stat function for every file. I don't know why the stat function is spent lots of time when the dictionary contains 25,000 files. I hope someone can help me. GlusterFS vesion: glusterfs 4.1.6 Volume info: Volume Name: gv0 Type: Distribute Volume ID: 7cfccb92-5b9d-4483-8212-0f02cd1197d6 Status: Started Snapshot Count: 0 Number of Bricks: 1 Transport-type: tcp Bricks: Brick1: dlaas-184:/data/glusterFS/gv0 Options Reconfigured: transport.address-family: inet nfs.disable: on
This bug is moved to https://github.com/gluster/glusterfs/issues/905, and will be tracked there from now on. Visit GitHub issues URL for further details