In that case, the trim/discard requests would need to come directly from the guest virtual machines to avoid damaging the filesystems. We do have a backlog feature ticket [1] to allow an administrator to transparently sparsify a in-use image via the rbd CLI, but no work has been started on it yet. [1] http://tracker.ceph.com/issues/13706 On Thu, Mar 2, 2017 at 5:16 AM, Massimiliano Cuttini <max@xxxxxxxxxxxxx> wrote: > Thanks Jason, > > I need some further info, because I'm really worried about ruin my data. > On this pool I have only XEN virtual disks. > Did I have to run the command directly on the "pool" or on the "virtual > disks" ? > > I guess that I have to run it on the pool. > As Admin I don't have access to local filesystem of the customer's virtual > disk and neither I can temporarly mount it to trim them. > Are my assumptions right? > > Another info: did I need to umount the image from every device that is > actually using the image while I'm trimming it? > > Thanks, > Max > > > > Il 01/03/2017 20:11, Jason Dillaman ha scritto: >> >> You should be able to issue an fstrim against the filesystem on top of >> the nbd device or run blkdiscard against the raw device if you don't >> have a filesystem. >> >> On Wed, Mar 1, 2017 at 1:26 PM, Massimiliano Cuttini <max@xxxxxxxxxxxxx> >> wrote: >>> >>> Dear all, >>> >>> i use the rbd-nbd connector. >>> Is there a way to reclaim free space from rbd image using this component >>> or >>> not? >>> >>> >>> Thanks, >>> Max >>> >>> _______________________________________________ >>> 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
 |