[Bug 578536] Re: when stopped, automount orphans some mounts

James Page james.page at ubuntu.com
Thu Apr 12 08:21:03 UTC 2012


@Chris

I've re-uploaded to natty-proposed with a much more verbose changelog
entry.

-- 
You received this bug notification because you are a member of Ubuntu
Sponsors Team, which is subscribed to the bug report.
https://bugs.launchpad.net/bugs/578536

Title:
  when stopped, automount orphans some mounts

Status in “autofs5” package in Ubuntu:
  Fix Released
Status in “autofs5” source package in Lucid:
  Fix Committed
Status in “autofs5” source package in Maverick:
  Fix Committed
Status in “autofs5” source package in Natty:
  In Progress
Status in “autofs5” source package in Oneiric:
  Fix Released

Bug description:
  Binary package hint: autofs5

  I have an automount setup containing a large number of NIS-provided
  maps, all derived from an NIS-provided master map (yp:auto.master).
  When I stop automount using "stop autofs", I noticed that some mounts
  have been unmounted while a non-deterministic number of mounts have
  been orphaned and need to be cleaned up using "umount":

  adar at adar-lx:~$ sudo start autofs && sudo stop autofs && grep autofs /proc/mounts
  autofs start/running, process 13152
  autofs stop/waiting
  auto.home /exit14 autofs rw,relatime,fd=19,pgrp=13152,timeout=60,minproto=5,maxproto=5,indirect 0 0
  /etc/auto.build /build autofs rw,relatime,fd=6,pgrp=13152,timeout=300,minproto=5,maxproto=5,indirect 0 0
  auto.exit15 /exit15 autofs rw,relatime,fd=31,pgrp=13152,timeout=60,minproto=5,maxproto=5,indirect 0 0
  auto.scripts /scripts autofs rw,relatime,fd=73,pgrp=13152,timeout=60,minproto=5,maxproto=5,indirect 0 0

  adar at adar-lx:~$ sudo start autofs && sudo stop autofs && grep autofs /proc/mounts
  autofs start/running, process 13198
  autofs stop/waiting
  /etc/auto.build /build autofs rw,relatime,fd=6,pgrp=13198,timeout=300,minproto=5,maxproto=5,indirect 0 0
  auto.scripts /scripts autofs rw,relatime,fd=73,pgrp=13198,timeout=60,minproto=5,maxproto=5,indirect 0 0
  auto.mts-tlv /mts-tlv autofs rw,relatime,fd=67,pgrp=13198,timeout=60,minproto=5,maxproto=5,indirect 0 0
  auto.mts-sof /mts-sof autofs rw,relatime,fd=79,pgrp=13198,timeout=60,minproto=5,maxproto=5,indirect 0 0
  auto.mts-cm /mts-cm autofs rw,relatime,fd=85,pgrp=13198,timeout=60,minproto=5,maxproto=5,indirect 0 0

  adar at adar-lx:~$ sudo start autofs && sudo stop autofs && grep autofs /proc/mounts
  autofs start/running, process 13244
  autofs stop/waiting
  auto.mts-sof /mts-sof autofs rw,relatime,fd=79,pgrp=13244,timeout=60,minproto=5,maxproto=5,indirect 0 0
  auto.buildmap /bldmnt autofs rw,relatime,fd=49,pgrp=13244,timeout=60,minproto=5,maxproto=5,indirect 0 0
  auto.exit26home /exit26 autofs rw,relatime,fd=55,pgrp=13244,timeout=60,minproto=5,maxproto=5,indirect 0 0

  I suspect this is due to the transition to upstart but I'm not 100%
  sure.

  ## SRU JUSTIFICATION ##
  IMPACT: Running reload autofs on a release with this issue creates automount maps and can cause the automounter to hang. This will prevent logins via ssh for users (due to NFS mounted home directories).

  FIX: Patch cherry picked from upstream,
  https://bugs.launchpad.net/ubuntu/+source/autofs5/+bug/578536/comments/6

  PATCH: see diff in unapproved queue.

  REPRODUCE: See this bug description

  DISCUSSION: Issue was known and resolved upstream, and has been
  carried in later releases of Ubuntu.  The patch coming untouched from
  upstream indicates that the regression potential is low.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/autofs5/+bug/578536/+subscriptions



More information about the Ubuntu-sponsors mailing list