Description of problem:
Geo-replication setup with other user then root on slave node,
as described in documentation:
If you run pause or resume on geo-replication with wrong user,
other user then you setup, you will get pause/resume success even
if it does not work.
To make it work again you need to restart glusterd on all nodes.
Version-Release number of selected component (if applicable):
glusterfs 3.12.5 (all packages)
If geo-replication is setup with user geoaccount, run:
gluster volume geo-replication [volume_name] anotheruser@[main_slave_node]::[volume_name] pause
Pausing geo-replication session between [volume_name] & anotheruser@main_slave_node]::[volume_name] has been successful
If you create a snapshot you will get the error that you need to stop geo-replication first.
So geo-replication does not pause
Steps to Reproduce:
1. gluster volume geo-replication [volume_name] anotheruser@[main_slave_node]::[volume_name] pause
2. gluster snapshot create [snap_name] [volume_name]
Geo-replication returns success with exit code 0 even if wrong user or no user (root) is specified at pause/resume when geo-replication user is setup.
Return code != 0 and an error message saying that wrong user is specified
Have not tested with root user for geo-replication and specifying another user at pause/resume
REVIEW: https://review.gluster.org/19658 (georep : Pause/Resume of geo-replication with wrong user) posted (#1) for review on master by Sunny Kumar
REVISION POSTED: https://review.gluster.org/19658 (georep : Pause/Resume of geo-replication with wrong user) posted (#2) for review on master by Sunny Kumar
Fix is present in releases 4.0 onwards, and 3.12 has reached EOL, request reports to update to the latest release for the required fix.