Description of problem: 1. when trying to mount ceph-fuse in /etc/fstab, i used the following example id=myuser,conf=/etc/ceph/cluster.conf /mnt/ceph2 fuse.ceph defaults,_netdev 0 0 and when i ran "mount -a" I am getting error saying as: mount: can't find id=myuser I tried mentioning keyring path and conf file path, the still exists and fails to mount.
does myuser exist in cephx ? "Ensure to use the ID (for example, admin, not client.admin)" please provide the output of ls /etc/ceph/*.keyring
Check that the mount helper (mount.fuse.ceph) is installed (should be in /usr/sbin)
I have reproduced this problem. [ubuntu@magna011 ~]$ ls /etc/ceph/*.keyring /etc/ceph/ceph.client.admin.keyring [ubuntu@magna011 ~]$ ls /usr/sbin | grep mount mount.ceph mount.fuse.ceph [ubuntu@magna011 ~]$ cat /etc/fstab # # /etc/fstab # Created by anaconda on Tue Jul 26 18:17:52 2016 # # Accessible filesystems, by reference, are maintained under '/dev/disk' # See man pages fstab(5), findfs(8), mount(8) and/or blkid(8) for more info # UUID=cbee517f-a45b-4fad-bbe8-d0316a34a263 / ext4 defaults 1 1 id=admin,conf=/etc/ceph/ceph.conf /mnt/ceph2 fuse.ceph defaults,_netdev 0 0 [ubuntu@magna011 ~]$ sudo mount -a mount: can't find id=admin,conf=/etc/ceph/ceph.conf I suspect that there may be something else that needs to be installed. I did a yum install ceph-fuse and yum install fuse prior to this.
mount.fuse was missing in the previous test. I installed it, saw it there, but still got the same message.
Googling mount.fuse.ceph we find this thread: http://lists.ceph.com/pipermail/ceph-users-ceph.com/2015-July/003111.html which leads to this BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1248003 It's not clear whether that ticket would ever have made to to the right RHEL people. Hopefully someone other than myself knows who the right RHEL people are?
Running the 7.3 util-linux, I get: sudo mount -a 2016-07-27 21:49:11.988047 7f860e4c9ec0 -1 init, newargv = 0x7f86192942a0 newargc=13 ceph-fuse[30323]: starting ceph client this is with the line: id=admin,conf=/etc/ceph/ceph.conf /mnt fuse.ceph defaults,_netdev 0 0 added to /etc/fstab.
Thanks Warren, And Gregory, Warren has provided all the information needed, so let me know if there is anything else that I can provide.
adding needinfo for Harish's query.
John, what workaround should we prescribe for customers in this case?
We already require people to be on RHEL 7.3 to use the kernel client, so maybe just extend the same requirement to use of the fuse client. I'm not very expert on how RHEL does business, was there some reason the fix for this is a 7.3 thing instead of an update to 7.2? That would be the best thing from my point of view, to fix the bug in the releases where it's present (otherwise I suspect we will be fielding "why is my fstab not working" questions for some time).
If RHEL 7.2's util-linux cannot activate ceph-fuse mounts, I'm confused how Teuthology was ever able to test ceph-fuse on CentOS 7.2? How are customers are expected to work around this issue until RHEL 7.3 ships? Is there some alternate way to mount ceph-fuse at boot time? (I'm hesitant to cross-ship the new util-linux in RHCS 2.0 this close to the release, because it's not clear how much testing this build has undergone, and regressions in this package could cause a system to fail to boot entirely)
and just going through the bz comments, i am pasting the util-linux used by teuthology runs if that helps. [ubuntu@clara010 ~]$ rpm -qa | grep util-linu util-linux-2.23.2-26.el7_2.2.x86_64
The note isn't quite accurate, it's not the kernel that is at fault here, it's the util-linux package. However I don't know if that will really confuse anyone, so I don't feel strongly.
Moving this bug to verified state in RHEL 7.3. Steps followed: 1. Configured ceph, mds and mounted cephfs to a client. 2. Added below entry in /etc/fstab/ "id=admin,conf=/etc/ceph/ceph.conf /mnt/myfs fuse.ceph defaults,_netdev 0 0" 3. executed "sudo mount -a" [ubuntu@node0 ~]$ sudo mount -a 2016-10-13 12:35:46.093350 7ff9d5ca9ec0 -1 init, newargv = 0x7ff9e1cdc2a0 newargc=13 ceph-fuse[2493]: starting ceph client ceph-fuse[2493]: starting fuse [ubuntu@node0 ~]$ 4. rebooted the client. 5. After reboot cephfs mount was available and able to run IO's.
Looks good to me, thanks
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHSA-2016-2815.html