Created attachment 1665166 [details] glusterfs-server cpu load graph Description of problem: Our glusterfs-volume(3.12.15) consists of 3 replications as shown below. Volume Name: nas Type: Distributed-Replicate Volume ID: 5a8cc386-85dc-4255-b287-352a499a28d5 Status: Started Snapshot Count: 0 Number of Bricks: 2 x 3 = 6 Transport-type: tcp Bricks: Brick1: NAS01:/gluster1/nas Brick2: NAS02:/gluster1/nas Brick3: NAS03:/gluster1/nas Brick4: NAS01:/gluster2/nas Brick5: NAS02:/gluster2/nas Brick6: NAS03:/gluster2/nas The current issue is the alternating cpu iowait and cpu load on three glusterfs-servers, with one, two or three simultaneously rising. It was confirmed that iowait rises first before the load goes up. However, since it is a volume consisting of 3 replications, the amount of reads and writes is similar, making it difficult to analyze the cause. The issue started from the day I remounted the volume on two clients (February 5), and I tried to remount and reboot the glusterfs-server to solve the problem, but to no avail. Can someone give me an answer? Version-Release number of selected component (if applicable): 3.12.15 How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
This bug is moved to https://github.com/gluster/glusterfs/issues/878, and will be tracked there from now on. Visit GitHub issues URL for further details