Bug 748888

Summary: -hosts auto-creates all available exports in /net but then can't make use any
Product: [Fedora] Fedora Reporter: Sandro Mathys <sandro>
Component: autofsAssignee: Ian Kent <ikent>
Status: CLOSED DUPLICATE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 16CC: ikent, lpoetter
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-10-26 06:58:51 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Sandro Mathys 2011-10-25 14:18:16 UTC 
With the standard /net -hosts definition, automounting NFS doesn't work. The very same works with F15. Exists in F16 TC2 (and probably earlier) both with and without latest from updates-testing.

If you try ls /net/your-nfsserver.localdomain/ it will create all available exports. If you try to enter one specific mount point, that will fail and the log says that's because 'file exists'.

[root@localhost ~]# mount | grep "/net"
cgroup on /sys/fs/cgroup/net_cls type cgroup (rw,nosuid,nodev,noexec,relatime,net_cls)
-hosts on /net type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,indirect)

[root@localhost ~]# ls /net/your-nfsserver.localdomain/
ID  MAVT  nsx1_watch_server_5

[root@localhost ~]# mount | grep "/net"
cgroup on /sys/fs/cgroup/net_cls type cgroup (rw,nosuid,nodev,noexec,relatime,net_cls)
-hosts on /net type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,indirect)
-hosts on /net/your-nfsserver.localdomain/ID/api/rhn/fs1/koji_db type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/your-nfsserver.localdomain/ID/api/rhn/fs1/koji_repo type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/your-nfsserver.localdomain/ID/api/rhn/fs1/koji_vcs type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/your-nfsserver.localdomain/ID/api/rhn/fs1/quota type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/your-nfsserver.localdomain/ID/api/stud/fs1/quota type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/your-nfsserver.localdomain/ID/api/stud/fs1/studroom type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/your-nfsserver.localdomain/ID/bd/db1/fs1/quota type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/your-nfsserver.localdomain/ID/bd/fs01/inst type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/your-nfsserver.localdomain/ID/bd/wcms/fs01/quota type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offse
-hosts on /net/your-nfsserver.localdomain/MAVT/iet/lsm/fs1/quota type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)
-hosts on /net/your-nfsserver.localdomain/nsx1_watch_server_5 type autofs (rw,relatime,fd=12,pgrp=2889,timeout=300,minproto=5,maxproto=5,offset)

[root@localhost ~]# grep automount /var/log/messages
Oct 25 16:12:16 localhost automount[2889]: move_mount: failed to move mount from /tmp/autoBlIxE5 to /net/your-nfsserver.localdomain/ID/api/stud/fs1/quota: File exists

[root@localhost ~]# rpm -qa autofs
autofs-5.0.6-2.fc16.x86_64
Comment 1 Ian Kent 2011-10-25 15:58:52 UTC 
This is because systemd marks the root file system shared which
causes any move mount system calls against any file systems to
fail.

Not really sure what to do about it and I'm not sure this is a
sensible thing for systemd to do either.

I guess that if you don't use the sandbox facility (at least I
think that's what it's called) you could probably mark the root
filesystem as private as a workaround.
Comment 2 Sandro Mathys 2011-10-26 06:17:23 UTC 
So, is there any plan to fix this? Is it an autofs or systemd issue, i.e. should I change the component this bug is reported against? Or shall we add systemd's maintainer to the CC list and ask for advice?

Also, how do you mark filesystems shared/private?
Comment 3 Sandro Mathys 2011-10-26 06:58:51 UTC 

*** This bug has been marked as a duplicate of bug 745781 ***