Hi Anny, On Thu, 2011-07-07 at 17:35 +0800, AnnyRen wrote: > Hi all, > > I have one questions about osd degraded. > > My environment was set up with 3 mon, 2 mds and 25 osd, all daemon are > running on physical machines. > Ceph Version: 0.30 > File system format: EXT4 > > Q1. When I'm testing osd stability, I kill an osd daemon (osd2) to > observe the behavior of data migration. > For all I know, I could see the pg degraded ratio down to zero, > and all pgs become active+clean. > but Ceph file system didn't migrate the data on the osd (which I > already killed its daemon) to other osds. > The degraded ratio still remain at 4.xxx%. > After then, I run "cosd -i 2 -c /etc/ceph/ceph.conf"and then run "ceph -w", > I saw all pgs suddenly become active+clean. > > Could anyone tell me why Ceph didn't migrate data when I kill one osd daemon? How long did you wait? The default configuration is that a OSD has to be down for at least 300 sec before it gets marked "out". When you kill the OSD it will be marked "down" within a few seconds, but that won't trigger any data migration, a OSD has to be marked "out" do to that. Wido > > > Thanks a lot~ ^__^ > > Best Regards, > Anny -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html