Description of problem: We lack cluster power down and reboot procedures in the Ceph Admin guide. It has been requested to include these procedures. Need to include the following components: - Monitors - OSDs - RGWs - MDSs - iSCSI GWs - Admin node (Ceph Metrics and Ansible) Refer: https://access.redhat.com/solutions/2139301 Version-Release number of selected component (if applicable): 3.0
The power down/up procedures for ceph are documented here: https://access.redhat.com/solutions/2139301 #Like Mike mentioned. Note: Step. 1 mentions clients... this includes MDSs. RGW/MDS order does not matter, but deactivation of MDS does covered next. PR: https://github.com/ceph/ceph/pull/21265 documents the current process for MDS take down. Unfortunately, before it had a chance to make it into RH documentation... behavior was enhanced upstream. The listed PR applies today(RHCS 3.0), and with 3.1 it "should" be simplified to 1 command for stop/start. REF(future release): - http://docs.ceph.com/docs/master/cephfs/administration/#taking-the-cluster-down - Current vs. New in 3.1: https://github.com/ceph/ceph/pull/16608/files#diff-2001b94c3deda3525a9380b7f3491875 <-- This is the doc diff... The PR has all changes which will show where we're headed FYA.