Re: RBD lock on unmount

David Turner <drakonstein@xxxxxxxxx> · Tue, 29 May 2018 07:32:12 -0400

This is not common unless you are using the kernel driver to map an RBD on a host running OSDs.  I've never had a problem unmounting an RBD that didn't have open file handles.  Note that Linux considers an FS to be active if a terminal, screen, etc is cd'd into the mounted directory.  Do you need to service files via NFS to non-linux servers?  If not, you might want to go the CephFS route.
If however you were using the kernel driver to map an RBD on a host running OSDs, then there is a known deadlock that can happen and try rebooting the machine before either changing the host you're mapping the RBD to or switching to rbd-fuse or rbd-nbd.

On Sun, May 27, 2018 at 8:13 PM Joshua Collins <joshua.collins@xxxxxxxxxx> wrote:

    Hi
    I've set up a Ceph Cluster with and RBD storage device for use in
      a pacemaker/corosync cluster. When attempting to move the
      resources from one node to the other, the filesystem on the RBD
      will not unmount. lsof and fuser show no files in use on the
      device. I thought this may be an issue with an NFS lock, so I
      moved the ceph OSD and monitor off the machine where the
      filesystem is mounted to a virtual machine, however I'm still
      unable to unmount the filesystem.
    Is this a known issue with RBD filesystem mounts? Is there a
      system change I need to make in order to get the filesystem to
      reliably unmount?
    Thanks in advance,
    -- 

                      Joshua Collins

                      Systems Engineer

                      38b Douglas Street

                      Milton QLD 4064

                      T  +61
                      7 3535 9615

                      F  +61
                      7 3535 9699

                      E  joshua.collins@xxxxxxxxxx

                      www.vrt.com.au

                      Follow us on:

                      This email is the property of VRT Systems, a
                      division of Vector International Pacific Pty. Ltd.
                      ACN: 009 278 979. The information contained in
                      this email is confidential and is intended only
                      for the use of the addressee(s). If you have
                      received this email in error, please immediately
                      notify the sender and delete the email. Thank you.

_______________________________________________

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