Hi, As you create your journal partition manually, you need to set a proper ceph partition GUID for the partition to be configured with correct ownership (ie ceph).
The issue has been faced by others before (including me), please refer to previous threads on the mailing list [0] [1] and you should be fine. [0] http://thread.gmane.org/gmane.comp.file-systems.ceph.user/27856 [1] http://thread.gmane.org/gmane.comp.file-systems.ceph.user/25734 On Fri, Apr 1, 2016 at 10:22 PM, Andrus, Brian Contractor <[email protected]> wrote: > All, > > > > I am trying to use ceph-deploy to create an OSD on a multipath device but > put the journal partition on the SSD the system boots from. > > I have created a partition on the SSD (/dev/sda5) but ceph-deploy does not > seem to like it. > > > > I am trying: > > *ceph-deploy osd create ceph01:/dev/mapper/mpathb:/dev/sda5* > > > > It runs through and ends with: > > [INFO ] Running command: ceph --cluster=ceph osd stat --format=json > > [WARNIN] there is 1 OSD down > > [WARNIN] there is 1 OSD out > > > > > > So I try to activate it: > > * ceph-deploy disk activate > ceph-1-35a:/dev/mapper/mpathb1:/dev/sda5* > > > > But that gives me an error: > > [WARNIN] 2016-04-01 13:07:55.519488 7fecdd044800 -1 ** ERROR: error > creating empty object store in /var/lib/ceph/tmp/mnt.VTxcYn: (13) > Permission denied > > > > It works fine if I only specify the single multipath device. > > It works fine if I point to a disk for the journal, but not a partition. > > It seems ceph-deploy is not set up to use an existing partition on a > separate disk for a journal? > > > > Is there a way to get ceph-deploy to use an existing partition for a > journal? > > > > Brian Andrus > > ITACS/Research Computing > > Naval Postgraduate School > > Monterey, California > > voice: 831-656-6238 > > > > _______________________________________________ > ceph-users mailing list > [email protected] > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > >
_______________________________________________ ceph-users mailing list [email protected] http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
