Bug#807769: systemd: /tmp mounted as tmpfs without user asking for it

[Sanjoy Mahajan]

Package: systemd
Version: 228-2
Severity: normal

-- Package-specific info:
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: not available
URL: <http://lists.alioth.debian.org/pipermail/pkg-systemd-maintainers/attachments/20151212/21717ba1/attachment.ksh>
-------------- next part --------------

   * What led up to the situation?

A bunch of reasonably important iceweasel-downloaded .pdf files in /tmp
suddenly disappeared.  I feared that I had misconfigured the
/etc/tmpfiles.d/tmp.conf and allowed automatic cleaning, but that file
was correct (prevented /tmp cleaning):

  # Automatically migrated from TMPTIME in /etc/default/rcS
  #d /var/tmp 1777 root root -
  d /tmp 1777 root root -

The syslog at around the time of the oldest remaining /tmp file showed
what had instead happened:

  Dec 11 07:24:57 insight dbus[671]: [system] Activating via systemd: service name='org.freedesktop.hostname1' unit='dbus-org.freedesktop.hostname1.service'
  Dec 11 07:24:57 insight systemd[1]: tmp.mount: Directory /tmp to mount over is not empty, mounting anyway.
  Dec 11 07:24:57 insight systemd[1]: Mounting Temporary Directory...
  Dec 11 07:24:57 insight systemd[1]: Mounted Temporary Directory.
  Dec 11 07:24:57 insight systemd[1]: Starting Hostname Service...
  Dec 11 07:24:57 insight dbus[671]: [system] Successfully activated service 'org.freedesktop.hostname1'
  Dec 11 07:24:57 insight systemd[1]: Started Hostname Service.

I had run 'hostnamectl' at around 7:24am, which activated the
org.freedesktop.hostname1 service, which then, via the 'PrivateTmp=yes'
line in dbus-org.freedesktop.hostname1.service used tmp.mount, which put
/tmp on tmpfs.

I got back my /tmp files by unmounting /tmp, so all was well.  But how
should I prevent it from happening again?  I don't want /tmp on tmpfs.
Should I just set PrivateTmp=no in that config?

I read bug #779902 and the patch in its msg #32
<https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=779902#32>, but I am
still confused about whether setting it to No has other bad effects.

*** End of the template - remove these template lines ***


-- System Information:
Debian Release: stretch/sid
  APT prefers testing
  APT policy: (990, 'testing'), (500, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.2.0-1-amd64 (SMP w/4 CPU cores)
Locale: LANG=en_US.utf8, LC_CTYPE=en_US.utf8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages systemd depends on:
ii  adduser         3.113+nmu3
ii  libacl1         2.2.52-2
ii  libapparmor1    2.10-2+b1
ii  libaudit1       1:2.4.4-4
ii  libblkid1       2.27.1-1
ii  libc6           2.19-22
ii  libcap2         1:2.24-12
ii  libcap2-bin     1:2.24-12
ii  libcryptsetup4  2:1.6.6-5
ii  libgcrypt20     1.6.4-3
ii  libkmod2        21-1
ii  liblzma5        5.1.1alpha+20120614-2.1
ii  libmount1       2.27.1-1
ii  libpam0g        1.1.8-3.1
ii  libseccomp2     2.2.3-2
ii  libselinux1     2.4-3
ii  libsystemd0     228-2
ii  mount           2.27.1-1
ii  sysv-rc         2.88dsf-59.2
ii  util-linux      2.27.1-1

Versions of packages systemd recommends:
ii  dbus            1.10.4-1
ii  libpam-systemd  228-2

Versions of packages systemd suggests:
pn  systemd-container  <none>
pn  systemd-ui         <none>

Versions of packages systemd is related to:
ii  udev  228-2

-- no debconf information