Intersting question. AFAIK, there's no built-in solution. Also, if you think about it, restarting whole cluster at once can lead to service interruption, as you can easily bring all PG copies in stale state for a short time, and even longer if some OSDs won't go up for some reason. You should really restart only single failure domain(e.g. node, rack, row, DC, or whatever your chrush ruleset assumes a failure domain to hold a separate PG copy) at the same time. Furthermore, there's rarely a need to restart OSD process, as you can change most of the configuration on the fly by injecting appropriate arguments to OSD/MON.
My best guess is that using systemd you can write some basic script to restart whatever OSDs you want. Another option is to use the same mechanics that ceph-deploy uses, but the principle is all the same - write some automation script. I would love to hear from people who has 100-1000+ OSDs in production though. 2017-05-12 15:53 GMT+03:00 Алексей Усов <[email protected]>: > Greetings, > > Could someone, please, tell me how do I restart all daemons in a cluster > if I make changes in ceph.conf, if it's needed indeed? Since > enterprise-scale ceph clusters usually tend to comprise of hundreds of > OSDs, I doubt one must restart the entire cluster by hand or use some sort > of external orchestrating tool - there must be a centralized solution in > ceph itself. Thank you in advance. > > -- > С уважением, Усов А.Е.. > Best Regards, Usov A.Y.. > > _______________________________________________ > ceph-users mailing list > [email protected] > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > >
_______________________________________________ ceph-users mailing list [email protected] http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
