Your message dated Mon, 18 Jan 2010 01:32:05 -0800
with message-id <[email protected]>
and subject line Re: Bug#565624: pam_limits: pam_limits should have a
non-infinite default for "as"
has caused the Debian Bug report #565624,
regarding pam_limits: pam_limits should have a non-infinite default for "as"
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 this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [email protected]
immediately.)
--
565624: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=565624
Debian Bug Tracking System
Contact [email protected] with problems
--- Begin Message ---
Package: libpam-modules
Version: 1.1.0-4
Severity: wishlist
File: pam_limits
In practice, it is silly not to have a default limit
for the maximum address space of a process. Once a
process exceeds the size of the RAM, things slow down
dramatically. Once a process exceeds RAM + swap space,
the system is toast.
The trouble is, that if a process starts running away,
allocating too much memory, then the computer sits there
thrashing, and it's often impossible to log in (if the
OOM killer kills sshd) or painfully slow. In such a
situation, it is not unusual for minutes to pass before
a shell prompt appears, and the only reasonable solutions
are to either kill the offending process or hit the
power button.
Far better to kill the runaway process before the
situation gets that bad. So, there should be a default
address space limit.
>From experiment, setting a limit at
min(RAM*1.2, RAM+0.5*swapspace)
is a good place. At that point, the system
is (almost always) thrashing, but (usually) not
badly enough as to force a shutdown. This is a
fairly conservative limit: often you'd want to
kill the process much earlier.
Someone who didn't like the limit could
easily over-ride it in /etc/default/limits.conf .
Note that this only handles the case where a single process
runs away. Obviously, if you had 3 or 4 identical processes,
the system could run out of memory much earlier. So, it
only a partial solution, but it will handle a common case.
-- System Information:
Debian Release: squeeze/sid
APT prefers testing
APT policy: (990, 'testing'), (500, 'testing-proposed-updates')
Architecture: i386 (i686)
Kernel: Linux 2.6.30-2-686 (SMP w/1 CPU core)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Versions of packages libpam-modules depends on:
ii debconf [debconf-2.0] 1.5.28 Debian configuration management sy
ii libc6 2.10.2-2 GNU C Library: Shared libraries
ii libdb4.7 4.7.25-8 Berkeley v4.7 Database Libraries [
ii libpam0g 1.1.0-4 Pluggable Authentication Modules l
ii libselinux1 2.0.89-4 SELinux runtime shared libraries
libpam-modules recommends no packages.
libpam-modules suggests no packages.
-- debconf information:
libpam-modules/disable-screensaver:
--- End Message ---
--- Begin Message ---
On Sun, Jan 17, 2010 at 02:21:27PM +0000, Greg Kochanski wrote:
> In practice, it is silly not to have a default limit
> for the maximum address space of a process. Once a
> process exceeds the size of the RAM, things slow down
> dramatically. Once a process exceeds RAM + swap space,
> the system is toast.
> The trouble is, that if a process starts running away,
> allocating too much memory, then the computer sits there
> thrashing, and it's often impossible to log in (if the
> OOM killer kills sshd) or painfully slow. In such a
> situation, it is not unusual for minutes to pass before
> a shell prompt appears, and the only reasonable solutions
> are to either kill the offending process or hit the
> power button.
> Far better to kill the runaway process before the
> situation gets that bad. So, there should be a default
> address space limit.
pam_limits doesn't set the default limits policy, it only echoes the default
policy that's set in the kernel. If you don't like the kernel defaults,
take it up with the kernel maintainers.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/
[email protected] [email protected]
signature.asc
Description: Digital signature
--- End Message ---
