[systemd-bugs] [Bug 90907] New: systemd automatically unmounts fuse file systems with default names

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=90907

            Bug ID: 90907
           Summary: systemd automatically unmounts fuse file systems with
                    default names
           Product: systemd
           Version: unspecified
          Hardware: Other
                OS: Linux (All)
            Status: NEW
          Severity: normal
          Priority: medium
         Component: general
          Assignee: systemd-bugs at lists.freedesktop.org
          Reporter: jacobsa at google.com
        QA Contact: systemd-bugs at lists.freedesktop.org

The default name for a file system mounted using fusermount without setting the
fsname option is "/dev/fuse". This interacts poorly with systemd, which appears
to think that it should automatically unmount any such file system.

For example, [1] is a bug thread for a fuse library whose tests are affected by
this. They are flaky on Ubuntu 15.04, often seeing that the file system has
been unmounted out from underneath them. This looks like the following in
/var/syslog:

    Jun  8 06:39:41 ubuntu systemd[1]: Unit tmp-fusetest301566509.mount is
bound to inactive unit dev-fuse.device. Stopping, too.
    Jun  8 06:39:41 ubuntu systemd[1]: Unmounting /tmp/fusetest301566509...

If the tests are modified to set an fsname option, this behavior goes away.
That is a fine workaround, but given the default behavior of fusermount these
tests will not be the last to trip over this.

I am far from a systemd expert, so I'm not confident what the correct answer is
here. But at first glance it seems rude for systemd to be interfering with fuse
file systems with default names.

[1]: https://github.com/bazil/fuse/issues/89

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/systemd-bugs/attachments/20150609/fd41d527/attachment.html>