When OSD gets full no further writes are possible. You need to rebalance the OSDs first.
Jan > On 09 Oct 2015, at 10:38, 陈积 <che...@letv.com> wrote: > > Hi, > I have deploy a small ceph cluster on my computer,and do a test like below > Firstly ,I make the /dev/sdb full manually by put a 8.3GB file on /mnt/tmp > ,and then try to write a 300M file through striprados which is a ceph client > using libradosstriper. The result is this write failed. > > So,Does ceph has any way to make full use of the left space of other disks in > this case? > > > sunfch@node1 <mailto:sunfch@node1> tmp]$ df -h > Filesystem Size Used Avail Use% Mounted on > /dev/mapper/VolGroup-lv_root 27G 21G 4.5G 82% / > tmpfs 940M 88K 940M 1% /dev/shm > /dev/vda1 485M 40M 421M 9% /boot > /dev/sda 10G 1.9G 8.2G 19% /var/lib/ceph/osd/ceph-0 > /dev/sdb 10G 10G 7.3M 100% /var/lib/ceph/osd/ceph-1 > /dev/sdc 10G 1.8G 8.3G 18% /var/lib/ceph/osd/ceph-2 > /dev/sdd 9.8G 1.9G 7.9G 20% /var/lib/ceph/osd/ceph-3 > /dev/sde 9.8G 1.8G 8.0G 19% /var/lib/ceph/osd/ceph-4 > /dev/sdf 9.8G 1.8G 8.1G 18% /var/lib/ceph/osd/ceph-5 > /dev/sdb 10G 10G 7.3M 100% /mnt/tmp > > [sunfch@node1 <mailto:sunfch@node1> tmp]$ striprados -p chenji -u 300m 300M > set up a rados cluster object > connected to the rados cluster > created an ioctx for our pool > created a striper for our pool > > fail > _______________________________________________ > 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
