| Hi all, Just wanted to explain my experience on how to stop the whole cluster and change the IPs. First, we shut down the cluster with this procedure: 1. Stop the clients from using the RBD images/Rados Gateway on this cluster or any other clients. 2. The cluster must be in healthy state before proceeding. 3. Set the noout, norecover, norebalance, nobackfill, nodown and pause flags #ceph osd set noout #ceph osd set norecover #ceph osd set norebalance #ceph osd set nobackfill #ceph osd set nodown #ceph osd set pause 4. Stop all ceph services 4.1. First osd nodes one by one 4.2. Lastly monitor nodes one by one Now we extracted the monmap with 'ceph-mon -i {mon-id} --extract-monmap /tmp/monmap‘Followed this manual: http://docs.ceph.com/docs/hammer/rados/operations/add-or-rm-mons/#changing-a-monitor-s-ip-address-the-messy-way And imported the new monmap to each monitor (while they all were stopped), changed ceph.conf on all nodes with the new IPs (don’t forget the clients). The last step was to change the IP Config and the hosts files (in our case, again don’t forget the clients) and shutdown the nodes. In the new datacenter we started the nodes and everything came up as usual. (1. Power on the admin node) 2. Power on the monitor nodes 3. Power on the osd nodes 4. Wait for all the nodes to come up , Verify all the services are up and the connectivity is fine between the nodes. 5. Unset all the noout,norecover,noreblance, nobackfill, nodown and pause flags. #ceph osd unset noout #ceph osd unset norecover #ceph osd unset norebalance #ceph osd unset nobackfill #ceph osd unset nodown #ceph osd unset pause 6. Check and verify the cluster is in healthy state, Verify all the clients are able to access the cluster. I hope this helps someone for the future!
|
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
