I updated your ticket http://tracker.ceph.com/issues/18589 with what I think is going on here. Note that adding "mds allow r" grants read capabilities on *everything*, which you probably don't want. -Greg On Wed, Jan 18, 2017 at 9:35 AM, Huamin Chen <hchen@xxxxxxxxxx> wrote: > Adding "mds caps allow r" fixes the problem. > > On Wed, Jan 18, 2017 at 11:38 AM, Huamin Chen <hchen@xxxxxxxxxx> wrote: >> >> Hi >> >> I have trouble mounting a ceph fs sub directory. I am using ceph jewel. >> [root@rootfs-dev cephfs]# ceph version >> ceph version 10.2.5 (c461ee19ecbc0c5c330aca20f7392c9a00730367) >> >> >> First I got permission denied with the following caps: >> >> [root@rootfs-dev cephfs]# ceph auth get client.bar >> exported keyring for client.bar >> [client.bar] >> key = AQA+ln9Yfm6DKhAA10k7QkdkfIAKqmM6xeCsxA== >> caps mds = "allow rw path=/volumes/kubernetes/test1" >> caps mon = "allow *" >> caps osd = "allow rwx pool=fs_data" >> [root@rootfs-dev cephfs]# mount -t ceph >> 172.24.0.4:6789:/volumes/kubernetes/test1 /tmp/mnt -o >> name=bar,secret=AQA+ln9Yfm6DKhAA10k7QkdkfIAKqmM6xeCsxA== >> mount error 13 = Permission denied >> >> >> Then I modified caps, this time I got I/O error. >> >> [root@rootfs-dev cephfs]# ceph auth caps client.bar mon 'allow *' mds >> 'allow *' osd 'allow *' >> updated caps for client.bar >> [root@rootfs-dev cephfs]# mount -t ceph >> 172.24.0.4:6789:/volumes/kubernetes/test1 /tmp/mnt -o >> name=bar,secret=AQA+ln9Yfm6DKhAA10k7QkdkfIAKqmM6xeCsxA== >> mount error 5 = Input/output error >> >> >> However ceph-fuse can mount the sub directory, using the same credential. >> >> [root@rootfs-dev cephfs]# ceph-fuse /mnt/test --id=bar >> --client-mountpoint=/volumes/kubernetes/test1 >> ceph-fuse[41114]: starting ceph client >> 2017-01-18 16:30:07.400280 7f83f2f27ec0 -1 init, newargv = 0x7f83fd126780 >> newargc=11 >> ceph-fuse[41114]: starting fuse >> >> Help appreciated. >> >> Thanks >> >> Huamin > > -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
