In my case one server was also non-GPT installed and in
/usr/sbin/ceph-disk I've added line:
os.chmod(os.path.join(path,'journal'), 0777) after line 1926.
I know that it's very ugly and shouldn't be made on production, but I
had no time to search for proper way to fix this.
Regards
Michał Chybowski
Tiktalik.com
W dniu 15.01.2016 o 05:00, Stuart Longland pisze:
On 12/01/16 01:22, Stillwell, Bryan wrote:
Well, it seems I spoke to soon. Not sure what logic the udev rules use
to identify ceph journals, but it doesn't seem to pick up on the
journals in our case as after a reboot, those partitions are owned by
root:disk with permissions 0660.
This is handled by the UUIDs of the GPT partitions, and since you're using
MS-DOS partition tables it won't work correctly. I would recommend switching to
GPT partition tables if you can.
I'm not comfortable with switching from MS-DOS to GPT disklabels on a
running production server, and nothing in the Ceph docs at the time
mentioned that GPT was a requirement on the journal disks.
Switching to GPT isn't toggling an option, it'd require resizing
partitions (some of these are xfs; can't be shrunk to my knowledge) and
moving stuff around to allow for an additional EFI boot partition,
possibly changes to boot firmware settings and bootloaders too.
I'll look into a udev rule for our particular case and see how I go.
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
