Open Captd65 opened 1 month ago
We noticed the following feature. If you add a task to cron on the server where glusterfs client is installed (runs every 5 minutes), a script that reads the list of directories and files (ls -lh /data/glustervolume) from the mounted folder with glusterfs, then this problem with file permissions is not observed. Maybe there is some kind of timeout that affects the failure of the client's rights?
Here is a similar problem: https://access.redhat.com/discussions/6958025
When I updated to version 11.1, the problem with the rights remained. Then I turned off the option "Readdirp = No" in "/etc/fstab." And the rights to folders stopped flying!
The problem is similar as in this request: (https://github.com/gluter/glusterfs/issues/71
But it looks like it is present in the new version 11.1?
Description of problem:
Good afternoon! Problem description: Replica + 1 referee has been created. This volume is connected to two hosts via the glusterfs client. Periodically, rights to some files or folders disappear on one of the clients, defined as ????. After remounting the partition or rebooting the problematic host, the permissions are displayed correctly. The exact command to reproduce the issue: It is not yet clear how to reproduce this problem, it reproduces itself during the day.
The full output of the command that failed: