Description of problem: ========================= glusterd is expected to restart only those processes/translators which are not running. It is expected to leave the rest untouched. 'gluster volume <volname> start force' (which is also the documented workaround for BZ 1341934) in turn ends up restarting the scrub process in every node, resetting the statistics that it would have collected until then. Version-Release number of selected component (if applicable): ======================================================== 3.7.9-7 How reproducible: ================ Always Steps to Reproduce: ================== 1. Have a volume with bitrot enabled and note down the scrub pid 2. Execute a gluster volume start force and again note down the scrub pid Actual results: ================ The pid changes after volume start force is executed. 'gluster volume bitrot scrub status' is reset to 0. Expected results: ================= Scrub process, if it is running, should not be impacted by volume start force. This in turn would leave the scrub status as how it was before.