CephFS

[Gregory Farnum <gfarnum@xxxxxxxxxx>]

On Sunday, January 17, 2016, James Gallagher <james.np.gallagher@xxxxxxxxx> wrote:

Hi,

I'm looking to implement the CephFS on my Firefly release (v0.80) with an XFS native file system, but so far I'm having some difficulties. After following the ceph/qsg and creating a storage cluster, I have the following topology

admin node - mds/mon

                   osd1

                   osd2

Ceph health is OK, ceph -s shows

monmap e1: 1 mons at {node1=192.168.43.129:6789/0}, election epoch 2, quorum 0 node 1

msdmap e6: 1/1/1 up {0=node1=up:active}

osdmap e10: 2osds: 2 up, 2 in

active + clean and so on

However, unfortunately when I use the guide here:

http://docs.ceph.com/docs/master/cephfs/kernel/

and try the command

sudo mount -t ceph 192.168.43.129:6789:/ /mnt/mycephfs -o name=admin-node,secretfile=admin.secret

where admin-node=hostname of admin node and where admin.secret is the string taken from ceph.client.admin.keyring without the unnecessary bits

I then get:

mount: wrong fs type, bad option, bad superblock on 192.168.43.129:6789

missing codepage or helper program or ...

for several filesystems e.g. nfs cifs

need a /sbin./mount.type helper program

This leads me to believe that there is a problem with XFS, but this is supported with this version of Ceph so I don't really know anymore.

When I try the command

sudo mount -t ceph 192.168.43.129:6789:/ /mnt/mycephfs -o name=admin-node,secret={secretkey}

I get libceph: auth method x error -1

mount: permission denied

and when I try

sudo mount -t ceph 192.168.43.129:6789:/ /mnt/mycephfs

I get

no secret set ...

error -22 on auth protocol 2 init

then the whole mount: wrong fs type, bad option jargon again

Any ideas?

What kernel are you running? What's the content of your admin.secret file?

-Greg

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com