Description of problem: We use v3.6.9 and we have a deployment with 2 storage nodes with bricks and a third one acting as quorum. We reboot these nodes sequentially: quorum node, second storage node and first storage node. After each reboot we give some time to system to become stable again. We have noticed that after the last reboot, renaming of a file fails with error 116 and, from this point on, glusterfsd of the related volume prints the following error logs: [2016-07-02 02:04:06.443589] E [posix.c:2774:posix_writev] 0-voldata-posix: write failed: offset 2141980, Bad file descriptor [2016-07-02 02:04:06.443623] I [server-rpc-fops.c:1378:server_writev_cbk] 0-voldata-server: 187238: WRITEV 2 (12952f99-2acb-4f6f-b08a-8d24737b1b82) ==> (Bad file descriptor) These errors stop only after restart of gluster clients. Could you please check and provide some help? Version-Release number of selected component (if applicable): v3.6.9 How reproducible: Very often Steps to Reproduce: 1. Reboot nodes as described above 2. Try to rename a file on a mounted volume Actual results: Expected results: Additional info: FYI we have back-ported the following patches from 3.7 branch: http://review.gluster.org/#/c/10829/ http://review.gluster.org/#/c/13358/ http://review.gluster.org/#/c/14052/ http://review.gluster.org/#/c/12310/ http://review.gluster.org/#/c/13392/
Can you confirm whether patches mentioned above fixes the issue? Now, 3.6 only receives security and critical fixes, Please refer this link: https://www.mail-archive.com/gluster-users%40gluster.org/msg25848.html You may want to upgrade your gluster to get fixes for this issue.
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 1000 days