On Mon, Mar 12, 2018 at 9:54 AM, Fulvio Galeazzi <fulvio.galeazzi@xxxxxxx> wrote: > Hallo all, > I am not sure RBD discard is working in my setup, and I am asking for > your help. > (I searched this mailing list for related messages and found one by > Nathan Harper last 29th Jan 2018 "Debugging fstrim issues" which > however mentions trimming was masked by logging... so I am not 100% > sure of what is the expected result) > > I am on Ocata, and Ceph 10.2.10. > Followed the recipe: > https://www.sebastien-han.fr/blog/2015/02/02/openstack-and-ceph-rbd-discard/ > * setup Nova adding to /etc/nova/nova.conf > ... > [libvirt] > hw_disk_discard = unmap > ... > * decorated a CentOS image with hw_scsi_model=virtio--scsi,hw_disk_bus=scsi Is that just a typo for "hw_scsi_model"? > * created a VM with boot disk on Ceph (my default is ephemeral, > though), verified the XML shows my disk is scsi, > > > I see that commands: > rbd --cluster cephpa1 diff cinder-ceph/${theVol} | awk '{ SUM += $2 } END { > print SUM/1024/1024 " MB" }' ; rados --cluster cephpa1 -p cinder-ceph ls | > grep rbd_data.{whatever} | wc -l That's pretty old-school -- you can just use 'rbd du" now to calculate the disk usage. > show the size increases but does not decrease when I execute delete the > temporary file and execute > sudo fstrim -v / > > Am I missing something? Have you verified that your VM is indeed using virtio-scsi? Does blktrace show SCSI UNMAP operations being issued to the block device when you execute "fstrim"? > I do see that adding/removing files created with dd does not always result > in a global size increase, it is as if the dirty blocks are kept around and > reused. Is this the way discard is supposed to work? > > Thanks for your help! > > Fulvio > > > _______________________________________________ > ceph-users mailing list > ceph-users@xxxxxxxxxxxxxx > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > -- Jason _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |