On Fri, May 04, 2018 at 12:08:35AM PDT, Tracy Reed spake thusly: > I've been using ceph for nearly a year and one of the things I ran into > quite a while back was that it seems like ceph is placing copies of > objects on different OSDs but sometimes those OSDs can be on the same > host by default. Is that correct? I discovered this by taking down one > host and having some pgs become inactive.
Actually, this (admittedly ancient) document: https://jcftang.github.io/2012/09/06/going-from-replicating-across-osds-to-replicating-across-hosts-in-a-ceph-cluster/ says "As the default CRUSH map replicates across OSD’s I wanted to try replicating data across hosts just to see what would happen." This would seem to align with my experience as far as the default goes. However, this: http://docs.ceph.com/docs/master/rados/operations/crush-map/ says: "When you deploy OSDs they are automatically placed within the CRUSH map under a host node named with the hostname for the host they are running on. This, combined with the default CRUSH failure domain, ensures that replicas or erasure code shards are separated across hosts and a single host failure will not affect availability." How can I tell which way mine is configured? I could post the whole crushmap if necessary but it's a bit large to copy and paste. -- Tracy Reed http://tracyreed.org Digital signature attached for your safety.
signature.asc
Description: PGP signature
_______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
