> Op 28 maart 2017 om 16:52 schreef Gregory Farnum <gfarnum@xxxxxxxxxx>: > > > CephFS files are deleted asynchronously by the mds once there are no more > client references to the file (NOT when the file is unlinked -- that's not > how posix works). If the number of objects isn't going down after a while, > restarting your samba instance will probably do the trick. In this case he was talking about RBD I see, not CephFS. A TRIM/Discard will need to be run on the Samba server to tell the block layer that the free space can be given back to Ceph. Wido > -Greg > On Tue, Mar 28, 2017 at 6:59 AM Götz Reinicke - IT Koordinator < > goetz.reinicke@xxxxxxxxxxxxxxx> wrote: > > > Hi, may be I got something wrong or did not understend it yet in total. > > > > I have some pools and created some test rbd images which are mounted to > > a samba server. > > > > After the test I deleted all files from the on the samba server. > > > > But "ceph df detail" and "ceph -s" show still used space. > > > > The OSDs on the ceph osd nodes are also filled with data. > > > > My question: At what point will the still existing but not needed/used > > objects be removed? > > > > > > Thanks for feedback and suggestions and Kowtow . Götz > > > > > > > > _______________________________________________ > > 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 _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |