The emperor’s new clothes, or “This pair of pants has three legs”

I’ve been trying to find a netbook distribution that works on my slightly ancient Asus eeePC 4G.

[insert howls of frustration here]

So far, not a single one that I’ve tried worked at all, with the exception of Easy Peasy.

Problem is, I wanted to replace Easy Peasy since the latest version (1.6) doesn’t work too well with the ridiculously small screen of the 4G…

So far I’ve tried MeeGo, Ubuntu 10.04 netbook remix and Moblin… any suggestions, anyone?


Performance tweaking with tmpfs

Some of you might have heard of tmpfs.
For those who haven’t, here’s a link to wikipedia. In short, tmpfs is what amounts to be the good old trusted ram disk. How can this be useful to speed up your computer?

Simple enough, use it for temporary files. On unix systems, that usually means putting /tmp on a tmpfs. On openSUSE, all you need to do is putting this line in your /etc/fstab (after cleaning out /tmp or the mount won’t work):

none                 /tmp                 tmpfs      size=1024m,defaults                   0 0

That line creates a ramdisk of 1Gbyte under /tmp.

but… what if you restart your computer? Right, everything in /tmp will be gone. So, if you don’t want to lose your stuff in there, you’ll need this little script:

#     Template SUSE system startup script for example service/daemon FOO
#     Copyright (C) 1995--2005  Kurt Garloff, SUSE / Novell Inc.
#     This library is free software; you can redistribute it and/or modify it
#     under the terms of the GNU Lesser General Public License as published by
#     the Free Software Foundation; either version 2.1 of the License, or (at
#     your option) any later version.
#     This library is distributed in the hope that it will be useful, but
#     WITHOUT ANY WARRANTY; without even the implied warranty of
#     Lesser General Public License for more details.
#     You should have received a copy of the GNU Lesser General Public
#     License along with this library; if not, write to the Free Software
#     Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307,
#     USA.
# /etc/init.d/synctmpfs
#   and its symbolic link
# /(usr/)sbin/rcsynctmpfs
# Template system startup script for some example service/daemon synctmpfs
# LSB compatible service control script; see                                                                                              
# Note: This template uses functions rc_XXX defined in /etc/rc.status on                                                                                                 
# UnitedLinux/SUSE/Novell based Linux distributions. If you want to base your
# script on this template and ensure that it works on non UL based LSB
# compliant Linux distributions, you either have to provide the rc.status
# functions from UL or change the script to work without them.
# See skeleton.compat for a template that works with other distros as well.
# Provides:          synctmpfs
# Required-Start:    boot.localfs
# Should-Start:     
# Required-Stop:     boot.localfs
# Should-Stop:      
# Default-Start:     3 5
# Default-Stop:      0 1 2 6
# Short-Description: synctmpfs XYZ daemon providing ZYX
# Description:       Start synctmpfs to allow XY and provide YZ
#       continued on second line by '#<TAB>'
#       should contain enough info for the runlevel editor
#       to give admin some idea what this service does and
#       what it's needed for ...
#       (The Short-Description should already be a good hint.)
# Any extensions to the keywords given above should be preceeded by
# X-VendorTag- (X-UnitedLinux- X-SuSE- for us) according to LSB.
# Notes on Required-Start/Should-Start:
# * There are two different issues that are solved by Required-Start
#    and Should-Start
# (a) Hard dependencies: This is used by the runlevel editor to determine
#     which services absolutely need to be started to make the start of
#     this service make sense. Example: nfsserver should have
#     Required-Start: $portmap
#     Also, required services are started before the dependent ones.
#     The runlevel editor will warn about such missing hard dependencies
#     and suggest enabling. During system startup, you may expect an error,
#     if the dependency is not fulfilled.
# (b) Specifying the init script ordering, not real (hard) dependencies.
#     This is needed by insserv to determine which service should be
#     started first (and at a later stage what services can be started
#     in parallel). The tag Should-Start: is used for this.
#     It tells, that if a service is available, it should be started
#     before. If not, never mind.
# * When specifying hard dependencies or ordering requirements, you can
#   use names of services (contents of their Provides: section)
#   or pseudo names starting with a $. The following ones are available
#   according to LSB (1.1):
#       $local_fs               all local file systems are mounted
#                               (most services should need this!)
#       $remote_fs              all remote file systems are mounted
#                               (note that /usr may be remote, so
#                                many services should Require this!)
#       $syslog                 system logging facility up
#       $network                low level networking (eth card, ...)
#       $named                  hostname resolution available
#       $netdaemons             all network daemons are running
#   The $netdaemons pseudo service has been removed in LSB 1.2.
#   For now, we still offer it for backward compatibility.
#   These are new (LSB 1.2):
#       $time                   the system time has been set correctly
#       $portmap                SunRPC portmapping service available
#   UnitedLinux extensions:
#       $ALL                    indicates that a script should be inserted
#                               at the end
# * The services specified in the stop tags
#   (Required-Stop/Should-Stop)
#   specify which services need to be still running when this service
#   is shut down. Often the entries there are just copies or a subset
#   from the respective start tag.
# * Should-Start/Stop are now part of LSB as of 2.0,
#   formerly SUSE/Unitedlinux used X-UnitedLinux-Should-Start/-Stop.
#   insserv does support both variants.
# * X-UnitedLinux-Default-Enabled: yes/no is used at installation time
#   (%fillup_and_insserv macro in %post of many RPMs) to specify whether
#   a startup script should default to be enabled after installation.
#   It's not used by insserv.
# Note on runlevels:
# 0 - halt/poweroff                     6 - reboot
# 1 - single user                       2 - multiuser without network exported
# 3 - multiuser w/ network (text mode)  5 - multiuser w/ network and X11 (xdm)
# Note on script names:
# A registry has been set up to manage the init script namespace.
# Please use the names already registered or register one or use a
# vendor prefix.

# Check for missing binaries (stale symlinks should not happen)
# Note: Special treatment of stop for LSB conformance

test -x $synctmpfs_BIN || { echo "$synctmpfs_BIN not installed";
 if [ "$1" = "stop" ]; then exit 0;
 else exit 5; fi; }

# Source LSB init functions
# providing start_daemon, killproc, pidofproc,
# log_success_msg, log_failure_msg and log_warning_msg.
# This is currently not used by UnitedLinux based distributions and
# not needed for init scripts for UnitedLinux only. If it is used,
# the functions from rc.status should not be sourced or used.
#. /lib/lsb/init-functions

# Shell functions sourced from /etc/rc.status:
#      rc_check         check and set local and overall rc status
#      rc_status        check and set local and overall rc status
#      rc_status -v     be verbose in local rc status and clear it afterwards
#      rc_status -v -r  ditto and clear both the local and overall rc status
#      rc_status -s     display "skipped" and exit with status 3
#      rc_status -u     display "unused" and exit with status 3
#      rc_failed        set local and overall rc status to failed
#      rc_failed <num>  set local and overall rc status to <num>
#      rc_reset         clear both the local and overall rc status
#      rc_exit          exit appropriate to overall rc status
#      rc_active        checks whether a service is activated by symlinks
. /etc/rc.status

# Reset status of this service

# Return values acc. to LSB for all commands but status:
# 0       - success
# 1       - generic or unspecified error
# 2       - invalid or excess argument(s)
# 3       - unimplemented feature (e.g. "reload")
# 4       - user had insufficient privileges
# 5       - program is not installed
# 6       - program is not configured
# 7       - program is not running
# 8--199  - reserved (8--99 LSB, 100--149 distrib, 150--199 appl)
# Note that starting an already running service, stopping
# or restarting a not-running service as well as the restart
# with force-reload (in case signaling is not supported) are
# considered a success.

case "$1" in
 echo -n "Syncing to tmpfs "
 ## Start daemon with startproc(8). If this fails
 ## the return value is set appropriately by startproc.
 rsync -aAX /tmp-bak/ /tmp/
 # Remember status and be verbose
 rc_status -v
 echo -n "backing up tmpfs "
 ## Stop daemon with killproc(8) and if this fails
 ## killproc sets the return value according to LSB.
 mkdir -p /tmp-bak
 rsync -aAX /tmp/ /tmp-bak/

 # Remember status and be verbose
 rc_status -v
 echo "Usage: $0 {start|stop}"
 exit 1


activate it with this command:

chkconfig synctmpfs on

On boot this will copy /tmp-bak into /tmp after the filesystems have been mounted, and on shutdown it will copy /tmp into /tmp-bak.

Now you’ll notice a distinct increase in speed at starting KDE, for example. Another application would be your browser cache.


Upgrading to 11.4, an even riskier way.

Here’s a little followup to my previous post about how to upgrade to 11.4.

This one describes how to make the whole process even more risky. You have been warned.
It’s even possible to do this remotely through a ssh session. If you attempt to do that, do it from inside a screen session!

Actually, this first bit makes it a bit easier:



mkdir -p ${newrepodir}

cd ${repodir}

for repofile in *repo; do
 echo -n converting ${repofile} to ${newrepodir}/$(echo ${repofile}|sed -e 's/11.3/11.4/g') ... ;
 cat "${repofile}" | sed -e 's/11.3/11.4/g' > "${newrepodir}/$(echo ${repofile}|sed -e 's/11.3/11.4/g')" ;
 echo done.


This script creates a directory /etc/zypp/repos.d_11.4 with all your configured repos switched to their 11.4 counterparts.

After you run this script, all you have to do is move the old /etc/zypp/repos.d out of the way, rename the new one to /etc/zypp/repos.d and make sure that all the paths in the repo files exist. If they don’t, “zypp ref” is going to complain.

now, before you begin, run these two commands:

zypper cc --all
zypper ref

The first one clears all zypper cache folders, the second one refreshes all repositories (here’s where you’ll see if all paths are correct).
If you have a custom X11 setup (dualscreen or such), now would be the point to make a copy of your xorg.conf file.
Now you jump into the instructions in my previous post, right at #6.

Upgrade to openSUSE 11.4, the risky way

Here’s a quick and dirty HOWTO about upgrading from openSUSE 11.3 to 11.4, with all additional repos enabled…

  1. Switch to runlevel 3: as root, execute “init 3”
  2. Login as root
  3. Make a backup of all your repositories:
    cd /etc/zypp
    cp -r repos.d repos.d_11.3
  4. Go through all your .repo files in /etc/zypp/repos.d and replace “11.3” with “11.4”.
    Here’s a “before/after” example:

    [openSUSE 11.3 OSS]
    name=Haupt-Repository (OSS)


    [openSUSE 11.4 OSS]
    name=Haupt-Repository (OSS)
  5. Refresh your repository cache:
    zypper ref
  6. Upgrade zypper:
    zypper install zypper

    You have to pay attention on the list of conflicts. Generally the best of the proposed solutions is the one the would uninstall the least number of packages while still installing/upgrading what you want.

  7. Do the full upgrade:
    zypper dup

    Again, pay attention to any listed conflicts, and resolve them carefully. Again, the best of the proposed solution would be the one that installs what you want while uninstalling as little as possible.

  8. The scary moment:
    Reboot your system. If all went well you will be greeted by a working openSUSE 11.4 installation.
  9. Cleaning up afterwards:
    Login as root, and run the following command:

    LANG=CTYPE zypper search -si | grep "(System Packages)"| cut -d "|" -f 2

    That will list all installed packages that are not available from any of the configured repositories anymore.
    You might want to carefully remove them one by one with “zypper remove”. If you use “zypper remove -u” instead, you will also remove all dependencies that are not needed anymore. This can really break things.