]> git.apps.os.sepia.ceph.com Git - ceph.git/commit
mon: health not ok when up < in osds
authorSage Weil <sage@newdream.net>
Mon, 29 Aug 2011 18:41:24 +0000 (11:41 -0700)
committerSage Weil <sage@newdream.net>
Mon, 29 Aug 2011 18:46:16 +0000 (11:46 -0700)
commit66b6289a1b64d2d77171a1071c098720a276f524
treede4215f3a557f590a6186fe8c341b47ce3e8c7be
parent502cf0b239eeaafb2b70cdf6cb5bdd4ab8d8b664
mon: health not ok when up < in osds

We were warning if there were any not up or in osds.  Instead, warn if
there are any osds that are in but not up.  That means if a node fails
and successfully marks the node out and retracts onto remaining nodes, the
ceph cluster is healthy again.

Presumably the fact that the nodes failed should raise other alerts,
because those specific daemons/nodes are not healthy.

Signed-off-by: Sage Weil <sage@newdream.net>
src/mon/OSDMonitor.cc