Ok thank you, so is there a way to unset my quota ? or should I create a new pool and destroy the old one ?
Another question by the way :) , does this syntax work : mount -t ceph ip1:6789,ip2:6789,ip3:6789:/qcow2 /disks/ If I only want to mount my "qcow2" pool ? >On Tue, Nov 5, 2013 at 5:09 PM, NEVEU Stephane ><stephane.ne...@thalesgroup.com> wrote: > Ok, so this command only works with rbd ? > ceph osd pool set-quota poolname max_bytes xxxx > > What happen then if I've already set a quota on my pool then added my > data_pool to mds ? Will this "quota" simply become ineffective in the cephfs > context ? and I'll be able to write larger datas than my "quota" ? > No. OSDs still enforce the quota. If quota has been reached, sync write will return -ENOSPC, buffered write will lose data. Yan, Zheng > -----Message d'origine----- > De : Yan, Zheng [mailto:uker...@gmail.com] Envoyé : mardi 5 novembre > 2013 09:38 À : NEVEU Stephane Cc : ceph-users@lists.ceph.com Objet : > Re: [ceph-users] [Cephfs] Mounting a specific pool > > On Tue, Nov 5, 2013 at 4:05 PM, NEVEU Stephane > <stephane.ne...@thalesgroup.com> wrote: >> Hi all, >> >> >> >> I'm trying to test/figure out how cephfs works and my goal is to >> mount specific pools on different KVM hosts : >> >> Ceph osd pool create qcow2 10000 >> >> Ceph osd dump | grep qcow2 >> >> -> Pool 9 >> >> Ceph mds add_data_pool 9 >> >> I want now a 900Gb quota for my pool : >> >> Ceph osd pool set-quota qcow2 max_bytes 120795955200 >> >> Ok, now how can I verify the size in Gb of my pool (not the >> replication size >> 1,2,3 etc) ? >> >> >> >> On my KVM host (client): >> >> Mount -t ceph ip1:6789,ip2:6789,ip3:6789:/ /disks/ >> >> OK >> >> Cephsfs /disks/ show_layout >> >> Layout.data_pool: 0 >> >> Etc. >> >> Cephfs /disks/ set_layout -p 9 -u 4194304 -c 1 -s 4194304 >> >> Umount /disks/ && mount /disks/ >> >> Cephsfs /disks/ show_layout >> >> Layout.data_pool: 9 >> >> >> >> Great my layout is now 9 so my qcow2 pool but : >> >> Df-h | grep disks, shows the entire cluster size not only 900Gb why ? >> is it normal ? or am I doing something wrong ? > > cephfs does not support any type of quota, df always reports ntire cluster > size. > > Yan, Zheng > > >> >> >> >> Thank you for your help J >> >> >> >> >> >> >> _______________________________________________ >> ceph-users mailing list >> ceph-users@lists.ceph.com >> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com >> _______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
