Hello Rainer, most of the time you just install the newer versions an restart the old ones without having to worry about the sequence. Otherwise just use a management solution that helps you with any day-to-day operation including the complete software update part. Something like you can see in this video https://www.youtube.com/watch?v=Jrnzlylidjs. -- Martin Verges Managing director Mobile: +49 174 9335695 E-Mail: martin.verges@xxxxxxxx Chat: https://t.me/MartinVerges croit GmbH, Freseniusstr. 31h, 81247 Munich CEO: Martin Verges - VAT-ID: DE310638492 Com. register: Amtsgericht Munich HRB 231263 Web: https://croit.io YouTube: https://goo.gl/PGE1Bx Am Di., 4. Juni 2019 um 11:19 Uhr schrieb Rainer Krienke <krienke@xxxxxxxxxxxxxx>: > > I have a fresh ceph 14.2.1 cluster up and running based on Ubuntu 18.04. > It consists of 9 hosts (+1 admin host). The nine hosts have each 16 > ceph-osd daemons running, three in these nine hosts also have > a ceph-mon and a ceph-mgr daemon running. So three hosts are running > osd, mon and also mgr daemons. > > Now I am unsure about the right way to go for ceph upgrades and linux > system host updates. > > Ceph-Upgrade: > Reading the ceph upgrade docs I ask myself how a future upgrade say to > 14.2.2 should be performed correctly? The recommendation says to upgrade > first monitors, then osds etc... > > So what is the correct way to go in a mixed setup like mine? Following > the rules strictly would mean not to use ceph-deploy install, but > instead to log into the mon(/osd) hosts and then upgrade only the > ceph-mon package and restart this mon, and then do the same with the > other monitors/osd hosts. After all mons have been successfully upgraded > I should then continue with upgrading OSDs (ceph-osd package) on one > host and restart all osds on this host one after another or reboot the > whole host. Then proceed to the next osd-host. > > Is this the correct and best way to go? > > Linux system updates: > The second point I would like to hear your opinions about is how you > handle linux system updates? Since even a non ceph linux system package > update might break ceph or even stop the whole linux host from booting, > care has to be taken. So how do you handle this problem? Do you run host > upgrades only manually in a fixed sequence eg first on a osd/mon host > and if the update is successful, then run the linux system package > updates on the other hosts? Do you use another strategy? > > Thanks > Rainer > -- > Rainer Krienke, Uni Koblenz, Rechenzentrum, A22, Universitaetsstrasse 1 > 56070 Koblenz, Tel: +49261287 1312 Fax +49261287 100 1312 > Web: http://userpages.uni-koblenz.de/~krienke > PGP: http://userpages.uni-koblenz.de/~krienke/mypgp.html > _______________________________________________ > ceph-users mailing list > ceph-users@xxxxxxxxxxxxxx > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |