[Pkg-sysvinit-devel] Bug#344001: marked as done (initscripts: /run
tmpfs should impose strict size and inode limits)
Debian Bug Tracking System
owner at bugs.debian.org
Sun Jan 8 19:03:48 UTC 2006
Your message dated Sun, 08 Jan 2006 19:58:55 +0100
with message-id <43C160EF.2090702 at yahoo.co.uk>
and subject line Bug in experimental fixed in experimental
has caused the attached Bug report to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what I am
talking about this indicates a serious mail system misconfiguration
somewhere. Please contact me immediately.)
Debian bug tracking system administrator
(administrator, Debian Bugs database)
--------------------------------------
Received: (at submit) by bugs.debian.org; 19 Dec 2005 10:52:03 +0000
>From rleigh at debian.org Mon Dec 19 02:52:03 2005
Return-path: <rleigh at debian.org>
Received: from s2.ukfsn.org ([217.158.120.143] helo=mail.ukfsn.org)
by spohr.debian.org with esmtp (Exim 4.50)
id 1EoIcd-0006zY-FA
for submit at bugs.debian.org; Mon, 19 Dec 2005 02:52:03 -0800
Received: from hardknott.home.whinlatter.ukfsn.org (84-45-213-194.no-dns-yet.enta.net [84.45.213.194])
by mail.ukfsn.org (Postfix) with ESMTP
id B6ADCE6F65; Mon, 19 Dec 2005 10:50:47 +0000 (GMT)
Received: from rleigh by hardknott.home.whinlatter.ukfsn.org with local (Exim 4.60)
(envelope-from <rleigh at debian.org>)
id 1EoIcZ-0001St-Sl; Mon, 19 Dec 2005 10:51:59 +0000
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
From: Roger Leigh <rleigh at debian.org>
To: Debian Bug Tracking System <submit at bugs.debian.org>
Subject: initscripts: /run tmpfs should impose strict size and inode limits
Message-ID: <20051219105159.5383.95059.reportbug at hardknott.home.whinlatter.ukfsn.org>
X-Mailer: reportbug 3.18
Date: Mon, 19 Dec 2005 10:51:59 +0000
Delivered-To: submit at bugs.debian.org
X-Spam-Checker-Version: SpamAssassin 2.60-bugs.debian.org_2005_01_02
(1.212-2003-09-23-exp) on spohr.debian.org
X-Spam-Level:
X-Spam-Status: No, hits=-8.0 required=4.0 tests=BAYES_00,HAS_PACKAGE
autolearn=no version=2.60-bugs.debian.org_2005_01_02
Package: initscripts
Version: 2.86.ds1-7
Severity: normal
Tags: experimental
/etc/init.d/mountvirtfs mounts a tmpfs on /run, but does so without
specifying any limits (size=nn), which means it defaults to half the
physical memory in the system.
Whilt in most cases this is appropriate, /run is intended to be very
small, and should only be used by a small number of packages. With
the default size, it creates the opportunity for a denial of service
attack (by filling the fs, exhausting available virtual memory), and
also the opportunity for abuse by package maintainers; by imposing
strict limits (possibly even nr_inodes), any abuse will be quickly
spotted.
Now that there are at least two tmpfs filesystems mounted by default
(/dev/shm and /run), it might be necessary to be a bit stricter
about the size of /dev/shm as well, since together both can be as
big as all the available memory. It would be prudent to check the
total VM size before using any default size.
For the /run size/inode limit, I would suggest adding something
similar to /etc/default/tmpfs (possibly even an addition to this
file, for example RUN_TMPFS_SIZE and SHM_TMPFS_SIZE).
Regards,
Roger
-- System Information:
Debian Release: testing/unstable
APT prefers unstable
APT policy: (990, 'unstable')
Architecture: powerpc (ppc)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.14.4
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)
Versions of packages initscripts depends on:
ii coreutils 5.93-5 The GNU core utilities
ii debianutils 2.15.2 Miscellaneous utilities specific t
ii dpkg 1.13.11 package maintenance system for Deb
ii e2fsprogs 1.38+1.39-WIP-2005.12.10-1 ext2 file system utilities and lib
ii libc6 2.3.5-9 GNU C Library: Shared libraries an
ii lsb-base 3.0-12 Linux Standard Base 3.0 init scrip
ii util-linux 2.12r-2 Miscellaneous system utilities
initscripts recommends no packages.
-- no debconf information
---------------------------------------
Received: (at 344001-done) by bugs.debian.org; 8 Jan 2006 18:58:59 +0000
>From jdthood at yahoo.co.uk Sun Jan 08 10:58:59 2006
Return-path: <jdthood at yahoo.co.uk>
Received: from smtp-out1.tiscali.nl ([195.241.79.176])
by spohr.debian.org with esmtp (Exim 4.50)
id 1Evfko-00078p-RX; Sun, 08 Jan 2006 10:58:59 -0800
Received: from [82.171.132.56] (helo=82-171-132-56.dsl.ip.tiscali.nl)
by smtp-out1.tiscali.nl with esmtp (Tiscali http://www.tiscali.nl)
id 1Evfko-0007HS-7N; Sun, 08 Jan 2006 19:58:58 +0100
Received: from [127.0.0.1] (localhost [127.0.0.1])
by 82-171-132-56.dsl.ip.tiscali.nl (Postfix) with ESMTP id 40D99BFEA6;
Sun, 8 Jan 2006 19:58:56 +0100 (CET)
Message-ID: <43C160EF.2090702 at yahoo.co.uk>
Date: Sun, 08 Jan 2006 19:58:55 +0100
From: Thomas Hood <jdthood at yahoo.co.uk>
User-Agent: Mozilla Thunderbird 1.0.7 (X11/20051013)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: 344089-done at bugs.debian.org, 343993-done at bugs.debian.org,
346139-done at bugs.debian.org, 344001-done at bugs.debian.org,
345321-done at bugs.debian.org
Subject: Bug in experimental fixed in experimental
X-Enigmail-Version: 0.92.1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
X-Spam-Checker-Version: SpamAssassin 2.60-bugs.debian.org_2005_01_02
(1.212-2003-09-23-exp) on spohr.debian.org
X-Spam-Level:
X-Spam-Status: No, hits=-2.5 required=4.0 tests=BAYES_00,SUSPICIOUS_RECIPS
autolearn=no version=2.60-bugs.debian.org_2005_01_02
X-CrossAssassin-Score: 2
This bug reported as occurring in an experimental version of sysvinit
and as fixed in a later experimental version of sysvinit is closed
with this message.
sysvinit (2.86.ds1-9) experimental; urgency=low
[ Thomas Hood ]
* Split mountdevsubfs out of mountvirtfs and run it at S04 and S37.
This should not behave any differently, but the split will allow
packages that futz with /dev to do this at S03.
* mountnfs.sh: Eliminate unnecessary use of subshell
* umountfs: Only run umount if there is something to unmount
(Closes: #345272)
* urandom: Set PATH so that find can be found (Closes: #345273)
* bootclean.sh: Improve, homogenize file deletion code
* init: 64_init_set_PATH.dpatch: Set PATH if it's unset on re-exec
(Closes: #345370)
* init: 65_init_u_in_06.dpatch: Allow 'telinit u' in runlevels 0, 6
(Closes: #345719)
* umountroot: Remount ro with -f on GNU/kFreeBSD (Closes: #344547)
* checkroot.sh, checkfs.sh: Pause for five seconds if sulogin fails
(Partially addresses: #337444)
* Include /var/log/fsck/ in initscripts package (Closes: #346139)
* Replace /lib/init/functions.sh with /lib/init/mount-functions.sh
for use by mountvirtfs and mountdevsubfs
* various initscripts: Clean up code that reads fstab
* various scripts: Redirect which program's stderr to /dev/null since
the GNU version prints an error message when the command is not
found (Closes: #345321)
* /etc/default/rcS: Make VERBOSE and DELAYLOGIN default to "no"
* /etc/default/rcS: Remove variable descriptions; refer to man page
* rcS.5: Fix descriptions of variables
* Remove obsolete docs
* Make initscripts Depend on mount >= 2.11x-1 (Closes: #345968)
* Remove obsolete dependencies on bsdutils, coreutils, dpkg, kbd and
util-linux: the versions in question are older than oldstable, so
any newly installed system and any system that has upgraded to
sarge (or even woody) satisfies the constraints.
* Remove obsolete dependency on the last package: last was forced off
systems when they upgraded to buzz
* Add Replaces to Conflicts: mdutils which no longer exists even in
oldstable
* initscripts postinst: Remove obsolete GMT-to-UTC code: this
conversion was performed when systems were upgraded to potato
* sysvinit: Depend on libc6 rather than Pre-Depending on it; we don't
do anything special in the preinsts any more
* Correct documentation of init's -e and -t options
* Tweak descriptions
* Add READMEs for remaining runlevel dirs (Closes: #242957)
* Previous release also
closes: #227540 "skeleton: Don't include /usr/local/* in PATH"
closes: #346415 "mountnfs.sh doesn't work with the "bg" mount option"
[ Petter Reinholdtsen ]
* sysv-rc: Fix some typos in the startpar handling. (Closes: #345269)
* Add code to detect and report bad 'exit' calls in init.d scripts,
if they kill /etc/init.d/rc.
* Make sure sourcing work with dash /bin/sh, by using 'set $action'
to pass arguments to the script. (Closes: #345267)
Thanks to Mark Hatle for help with this release.
-- Petter Reinholdtsen <pere at debian.org> Sun, 8 Jan 2006 18:08:51 +0100
sysvinit (2.86.ds1-8) experimental; urgency=low
[ Thomas Hood ]
* Omit /run until we are sure it's needed (Closes: #344001)
* Previous release also
closes: #338736 "mountvirtfs: Succeeds in various cases it should fail"
closes: #342160 "checkroot.sh ignores fsck result"
closes: #342744 "checkroot.sh sources mountvirtfs, yet attemps to supply arguments"
* Use /proc/mounts instead of /etc/mtab when unmounting (Closes: #338801)
* Override lintian warnings re: top-level dirs
* Check for files under mountpoints more thoroughly
* initscripts.postinst:
+ Don't fail to install on file-rc systems (Closes: #343993)
+ Don't fail to install in chroots (Closes: #344089)
* No longer keep the dynamic nologin flag file on the root filesystem;
instead, keep it at /var/lib/initscripts/nologin. Note to admins:
initscripts's postinst symlinks /etc/nologin to the latter location.
To switch login delaying on or off permanently, set DELAYLOGIN=no in
/etc/default/rcS and either create or delete (respectively)
/var/lib/initscripts/nologin.
* mountvirtfs: Mount /dev/shm earlier; move long comment to README.Debian
* Add 45_pidof_symlink.dpatch: Make pidof an absolute symlink
(Closes: #343862)
* debian/rules, 30_strip.dpatch: Strip .comment section from executables
(Closes: #343863)
* In umountfs, run umount with -f (Closes: #344547)
[ Petter Reinholdtsen ]
* Enable the startpar option as it should work now that .sh scripts
are serialized. The clock should no longer jump while startpar
is used.
-- Petter Reinholdtsen <pere at debian.org> Tue, 27 Dec 2005 14:18:19 +0100
--
Thomas Hood
More information about the Pkg-sysvinit-devel
mailing list