Unix Resources: July 2007

Tuesday, July 17, 2007

Configuring kvpnc

Here is the configuration options I use. Everything else should not be selected :

Profile
- General
-- Advanced
---- Enable advanced settings
---- Perfect Forward secrecy (PFS): server

- Authenticate
-- User data
---- Username : put your username
---- Password : put your password
---- Save user password
-- PSK
---- Save PSK
---- Pre shared key: Company pre-shared key

-- Network
--- General
---- Use connection status check
------ Interval : 1
------ Success count : 4
---- Use specified address to ping : put your workstation IP
--- Routes
---- Replace default route
--- NAT
---- Use UDP (NAT-T)

I sometime get timeout on the VPN server but If I retry it works.

Debugging LIRCD

Here are some information I took on a working machine :

lirc stuff works like this :
Kernel module gets IR stuff -> /dev/lirc
lircd reads /dev/lirc and uses /etc/lircd.conf to get /dev/lircd
irw attaches to /dev/lircd
mode2 attaches to /dev/lirc
problem is I think only one program can attach to /dev/lirc but gladly
lircd will accept multiple connections

[root@moon dev]# ls -l lirc*
lrwxrwxrwx 1 root root 5 Mar 18 10:11 lirc -> lirc0
crw------- 1 root root 61, 0 Mar 18 10:11 lirc0
srw-rw-rw- 1 root root 0 Mar 18 14:29 lircd
prw-r--r-- 1 root root 0 Mar 18 10:11 lircm

[root@moon dev]# lsmod | grep -i lirc
lirc_pvr150 19136 5
lirc_dev 12708 1 lirc_pvr150
ivtv 175760 5 lirc_pvr150
i2c_core 22209 10 lirc_pvr150,wm8775,cx25840,tda9887,tuner,ivtv,i2c_algo_bit,tveeprom,nvidia,i2c_i801

[root@moon dev]# mode2
code: 0x1794
code: 0x1f95
code: 0x1797
code: 0x1f96
code: 0x1f96

an interesting discussion can be found <a href="http://www.nabble.com/No--dev-lirc-when-upgrading-from-FC6-from-FC5.-t3167625.html">here</a>.

Compiling Cisco vpnclient-linux-x86_64-4.8 on Fedora Core 6 x86_64 kernel 2.6.18/2.6.19

Thanks a lot to Amit who found the following solution ( http://blog.360.yahoo.com/blog-.WURHFYwdq8.zfEosWC6j8jQ?p=55 )

Unfortunately the cisco vpnclient-linux-x86_64-4.8 will not compile with the kernel (2.6.18) which comes with fedora core 6 or 2.6.19(latest stable from http://kernel.org) both x86 and x86_64 .

 Error(s) you get while compiling

#./vpn_install
Cisco Systems VPN Client Version 4.8.00 (0490) Linux Installer
Copyright (C) 1998-2005 Cisco Systems, Inc. All Rights Reserved.
By installing this product you agree that you have read the
license.txt file (The VPN Client license) and will comply with
its terms.
Directory where binaries will be installed [/usr/local/bin]
Automatically start the VPN service at boot time [yes]
In order to build the VPN kernel module, you must have the
kernel headers for the version of the kernel you are running.
Directory containing linux kernel source code [/lib/modules/2.6.19-1.meaks/build]
* Binaries will be installed in "/usr/local/bin".
* Modules will be installed in "/lib/modules/2.6.19-1.meaks/CiscoVPN".
* The VPN service will be started AUTOMATICALLY at boot time.
* Kernel source from "/lib/modules/2.6.19-1.meaks/build" will be used to build the module.
Is the above correct [y]
Making module
make -C /lib/modules/2.6.19-1.meaks/build SUBDIRS=/home/amitkr/setups/vpnclient modules
make[1]: Entering directory `/home/amitkr/kernel/linux-2.6.19-1.meaks/build'
make[1]: Warning: File `Makefile' has modification time 1e+04 s in the future
make -C /home/amitkr/kernel/linux-2.6.19-1.meaks O=/home/amitkr/kernel/linux-2.6.19-1.meaks/build modules
CC [M] /home/amitkr/setups/vpnclient/interceptor.o
In file included from /home/amitkr/setups/vpnclient/Cniapi.h:15,
from /home/amitkr/setups/vpnclient/interceptor.c:30:
/home/amitkr/setups/vpnclient/GenDefs.h:110:2: warning: #warning 64 bit
/home/amitkr/setups/vpnclient/interceptor.c: In function handle_vpnup:
/home/amitkr/setups/vpnclient/interceptor.c:310: warning: assignment from incompatible pointer type
/home/amitkr/setups/vpnclient/interceptor.c:334: warning: assignment from incompatible pointer type
/home/amitkr/setups/vpnclient/interceptor.c:335: warning: assignment from incompatible pointer type
/home/amitkr/setups/vpnclient/interceptor.c: In function do_cleanup:
/home/amitkr/setups/vpnclient/interceptor.c:378: warning: assignment from incompatible pointer type
/home/amitkr/setups/vpnclient/interceptor.c: In function recv_ip_packet_handler:
/home/amitkr/setups/vpnclient/interceptor.c:553: error: CHECKSUM_HW undeclared (first use in this function)
/home/amitkr/setups/vpnclient/interceptor.c:553: error: (Each undeclared identifier is reported only once
/home/amitkr/setups/vpnclient/interceptor.c:553: error: for each function it appears in.)
/home/amitkr/setups/vpnclient/interceptor.c:557: error: too many arguments to function skb_checksum_help
/home/amitkr/setups/vpnclient/interceptor.c: In function do_cni_send:
/home/amitkr/setups/vpnclient/interceptor.c:680: error: CHECKSUM_HW undeclared (first use in this function)
/home/amitkr/setups/vpnclient/interceptor.c:683: error: too many arguments to function skb_checksum_help
make[4]: *** [/home/amitkr/setups/vpnclient/interceptor.o] Error 1
make[3]: *** [_module_/home/amitkr/setups/vpnclient] Error 2
make[2]: *** [modules] Error 2
make[1]: *** [modules] Error 2
make[1]: Leaving directory `/home/amitkr/kernel/linux-2.6.19-1.meaks/build'
make: *** [default] Error 2
Failed to make module "cisco_ipsec.ko".

 What is the Problem ?

This is because of
[1]
no more exists with, which used to be present in ${KSRCPATH}/build/include/linux/config.h with old kernels, after doing a make O=build menuconfig

[2]
For kernel 2.6.19 things are even worse, the CHECKSUM_HW macro does not exists any more in

[3]
the function skb_checksum_help(skb) declared in file and defined in file has been changed to take only a single argument.

 Solution

So, to compile the cisco vpn 4.8 x86_64 you need to:
[1] make a symlink to autoconfig.h as config.h
$ cd ${KSRCPATH}/build/include/linux
$ ln -s autoconfig.h config.h
This will solve the first problem

[2] the macro CHECKSUM_HW needs to be replaced with CHECKSUM_COMPLETE in the file ${VPNCLIENT}/interceptor.c

[3] edit the files calling skb_checksum_help() , look for the proper LINUX_VERSION_CODE macro that matches your kernel version(2.6.19) and remove the second argument from skb_checksum_help() so that the function looks like skb_checksum_help(skb).
The macro looks something like this

KERNEL_VERSION(2,6,10)

You can find the changes in concurrent releases of kernel in kernel ChangeLog

Create a dedicated user for mysql backup script

Here is how you create a backup user that will be used in backup scripts :

[root@MyServer /]# mysql -u root -p
Enter password:
Welcome to the MySQL monitor. Commands end with ; or \g.
Your MySQL connection id is 89061 to server version: 4.1.12

Type 'help;' or '\h' for help. Type '\c' to clear the buffer.

mysql> GRANT SELECT, SHOW DATABASES, LOCK TABLES ON *.* TO 'backup'@'localhost' IDENTIFIED BY 'MyPasswd';
Query OK, 0 rows affected (0.00 sec)

mysql> exit
Bye

Changing a user's passwd with mysql command line

Using PHPMyAdmin simplifies greatly MySQL database administration, but there are times where this software is not available...

So to change a user's password from the commandline, follow this procedure :
[root@MyServer /]# mysql -u root -p
Enter password:
Welcome to the MySQL monitor. Commands end with ; or \g.
Your MySQL connection id is 89057 to server version: 4.1.12

Type 'help;' or '\h' for help. Type '\c' to clear the buffer.

mysql> set password for 'root'@'localhost' = PASSWORD('MyNewPassword');
Query OK, 0 rows affected (0.05 sec)

Converting video_ts and audio_ts to an iso image

You have your DVD structure on your hard disk and you want to convert it to an iso image file, here is what you do :
mkisofs -pad -J -R -oYourDVD.iso -graft-points "/AUDIO_TS=/path/to/AUDIO_TS" "/VIDEO_TS=/path/to/VIDEO_TS"

Mount Bin/Cue files in Linux

It’s really easy to convert bin/cue files into an ISO file, and then mount the ISO in linux.

yum install bchunk

The syntax from bchunk is as follows:
bchunk [-v] [-p] [-r] [-w] [-s]

So if i wanted to convert image,bin and image.cue into image.iso, I’d run the command:
bchunk image.bin image.cue image.iso

Then to mount the ISO in linux you run the command:
mount -o loop -t iso9660 image.iso /mnt/image, where image.iso is the iso is the image that you want to mount and /mnt/image is the mount directory.

NX client and linux

In order to install nxclient for linux we need :
yum install compat-libstdc++-296

Setting up NFS on Solaris

Setup the server :

Edit /etc/dfs/dfstab :

share -F nfs -o root=client_nfs.domain.com -d "Backup FS" /raid5

Note : By default root is assigned nobody permission on an NFS share. To enable root permission on the share, we need to specify root= in the share? options.

Start the nfs server :

# /etc/init.d/nfs.server start

#

Check if the share is effective :

# showmount -e

export list for server_nfs:

/raid5 @204.19.29.0

#

To unshare :

# unshare /raid5

Setup the client :

mount server_nfs:/raid5 /nfs

Check the mount :

root@client_nfs:nfsstat -m

/nfs from server_nfs:/raid5

Flags: vers=3,proto=tcp,sec=sys,hard,intr,link,symlink,acl,rsize=32768,wsize=32768,retrans=5

root@client_nfs:

Setup automatic mount at boot time :

Edit /etc/vfstab and add the following line :

columbia:/raid5 - /nfs nfs - yes -

Availability tests

Case 1 :

Initial state : NFS Server is UP, NFS client has a mounted share.

Observation : Everything works perfectly.

Failure : NFS Server is DOWN, NFS client still has a mounted share.

Observation : NFS client still works fine. The share is not accessible. When we try to access it, the command froze. No implication on the system.

Recovery : NFS Server is brought back UP.

NFS client automatically finds the share. We can use it. Everything is automatic.

Case 2 :

Initial state : NFS Server is UP, NFS client has a mounted share.

Observation : Everything works perfectly.

Failure : Client is DOWN

Recovery : Client is brought back UP .

Observation : The client is able to use the share. Everything is back to normal.

Case 3 :

Initial state : NFS Server is UP, NFS client has a mounted share.

Observation : Everything works perfectly.

Failure : NFS Server AND NFS client shutdown due to a power loss.

Recovery : NFS client go back up first. NFS Server is down.

NFS client goes back UP and is ready. We can access the shared directory but nothing is there.

NFS Server is brought back UP. The NFS client automatically remount the share and can access it.

Everything is back to normal.

Monday, July 16, 2007

Preparing DVD from RHEL 4 installation CDs

Hi,

Ever wonder how to create a DVD from the 4 RHEL 4 installation CDs ? Well Chris Kloiber seems to have made up a script that works very well on Fedora Core 5.

Requirements :
The installation ISO CD image of RHEL (but I suspect this would work with any distributions)

Here is the script :

#/bin/bash

# by Chris Kloiber <ckloiber@redhat.com>

# A quick hack that will create a bootable DVD iso of a Red Hat Linux
# Distribution. Feed it either a directory containing the downloaded
# iso files of a distribution, or point it at a directory containing
# the "RedHat", "isolinux", and "images" directories.

# This version only works with "isolinux" based Red Hat Linux versions.

# Lots of disk space required to work, 3X the distribution size at least.

# GPL version 2 applies. No warranties, yadda, yadda. Have fun.

if [ $# -lt 2 ]; then
echo "Usage: `basename $0` source /destination/DVD.iso"
echo ""
echo " The 'source' can be either a directory containing a single"
echo " set of isos, or an exploded tree like an ftp site."
exit 1
fi

cleanup() {
[ ${LOOP:=/tmp/loop} = "/" ] && echo "LOOP mount point = \/, dying!" && exit
[ -d $LOOP ] && rm -rf $LOOP
[ ${DVD:=~/mkrhdvd} = "/" ] && echo "DVD data location is \/, dying!" && exit
[ -d $DVD ] && rm -rf $DVD
}

cleanup
mkdir -p $LOOP
mkdir -p $DVD

if [ !`ls $1/*.iso 2>&1>/dev/null ; echo $?` ]; then
echo "Found ISO CD images..."
CDS=`expr 0`
DISKS="1"

for f in `ls $1/*.iso`; do
mount -o loop $f $LOOP
cp -av $LOOP/* $DVD
if [ -f $LOOP/.discinfo ]; then
cp -av $LOOP/.discinfo $DVD
CDS=`expr $CDS + 1`
if [ $CDS != 1 ] ; then
DISKS=`echo ${DISKS},${CDS}`
fi
fi
umount $LOOP
done
if [ -e $DVD/.discinfo ]; then
awk '{ if ( NR == 4 ) { print disks } else { print ; } }' disks="$DISKS" $DVD/.discinfo > $DVD/.discinfo.new
mv $DVD/.discinfo.new $DVD/.discinfo
fi
else
echo "Found FTP-like tree..."
rsync -avP --exclude SRPMS $1/* $DVD
# cp -av $1/* $DVD
[ -e $1/.discinfo ] && cp -av $1/.discinfo $DVD
fi

rm -rf $DVD/isolinux/boot.cat
find $DVD -name TRANS.TBL | xargs rm -f

# My thanks to Mubashir Cheema for suggesting this fix.
# cd $DVD
mkisofs -J -R -v -T -o $2 -b isolinux/isolinux.bin -c isolinux/boot.cat -no-emul-boot -boot-load-size 8 -boot-info-table $DVD

/usr/lib/anaconda-runtime/implantisomd5 --force $2
# Don't like forced mediacheck? Try this instead.
# /usr/lib/anaconda-runtime/implantisomd5 --supported-iso --force $2

cleanup
echo ""
echo "Process Complete!"
echo ""

Monitoring RAID status

When using a RAID hardware on a linux server, we need to find out if a disk failed. If you use a Compaq/HP server, chances are that you are using a CCISS driver to access the raid device.

In that case, you can install an HP package that will work on fedora and other flavour of linux. You get it from <a href="wget ftp://ftp.compaq.com/pub/products/servers/supportsoftware/linux/hpacucli-7.20-16.linux.rpm">here</a>.

Then you can use the following 2 scripts :

Here is "hwraidinfo":
-----cut here----------
#!/bin/sh
SLOTLIST=$(hpacucli ctrl all show | \
grep Slot | sed -e 's/^.*Slot //g' -e 's/ .*$//g')

for i in $SLOTLIST
do
echo
hpacucli ctrl slot=$i show | grep -v "^$"
echo
hpacucli ctrl slot=$i ld all show | grep -v "^$"
hpacucli ctrl slot=$i pd all show | grep -v "^$"
done
echo
---------cut here--------------
And "hwraidstatus":
---------cut here--------------
#!/bin/sh
SLOTLIST=$(hpacucli ctrl all show | \
grep Slot | sed -e 's/^.*Slot //g' -e 's/ .*$//g')

for i in $SLOTLIST
do
echo
hpacucli ctrl slot=$i show status | grep -v "^$"
echo
hpacucli ctrl slot=$i ld all show status | grep -v "^$"
hpacucli ctrl slot=$i pd all show status | grep -v "^$"
done
echo
-------------cut here------------

This was posted by Matti Kurkela.

Getting VMware-player to work on Fedora Core 5

This is an extract from <a href="http://clunixchit.blogspot.com/2006_03_01_clunixchit_archive.html">this page</a>.

Getting VMware-player-1.0.1-19317.i386.rpm work on Fedora Core 5 'Bordeaux':
1. Get the rpm of the free VMWare Player
2. Download the linux image e.g VMware-player-1.0.1-19317.i386.rpm
3. Install it with rpm -ivh VMware-player-1.0.1-19317.i386.rpm
4. Then its time to download the patch: wget -c http://ftp.cvut.cz/vmware/vmware-any-any-update104.tar.gz
5. Extract the archive: tar xvf vmware-any-any-update104.tar.gz
6. cd vmware-any-any-update104
7. Complete the installation by: su -c "./runme.pl"

Adressing StorEdge L280 connected to SUN E450 running Solaris 9 in 64 bits mode with MTX and sst

<H3><BOLD>BEFORE YOU EVEN CONSIDER GETTING A SUN StorEdge L280 :</BOLD></H3>

Get the proper SUN Part. Be aware that the L280 won't connect through the standard SCSI port on the E450. It is single ended whereas the L280 requires a differential SCSI bus. Therefore you will need a differential SCSI host adapter (you can get it around 700 CAD$) and the proper cables. This is SUN option number X6541A (which is also SUN Part number 375-0006).

Introduction
When you buy a Sun StorEdge L280, it consists of a tape drive and a robotic device. You can address the tape drive as you do with standard tape drive (/dev/rmt). But you won't be able to address the robotic device. Therefore you will have to load tapes manually or buy a commercial backup software that will recognise and will be able to use the robot. 
There is a third solution which is the purpose of this document. You can use open source software to address the robotic device. and it works !

Preparing the installation
Power up your L280. Make sure the L280 is in random mode and write down your tape drive and your library SCSI IDs (you will need them later on). 
Begin with installing gcc3.2. You can get the package gcc-3.2-sol9-sparc-local.gz from <a href="http://www.sunfreeware.com">www.sunfreeware.com</a>. Note that this is actually a sun package. So, when you unzip it, type pkgadd -d gcc-3.2-sol9-sparc-local to install it. 
make sure your path includes /usr/local/bin:/usr/ccs/bin. 
To be able to address your library, you need a SCSI driver (it is not provided by SUN when you buy the StorEdge L280). This SCSI driver is sst. You can download it from <a href="http://www.arkeia.com/download.html">www.arkeia.com</a> and mtx software from <a href="http://mtx.badtux.net/">mtx.badtux.net</a>. 

Installing softwares and connecting the devices
untar the sst driver and follow the procedure describes in the sst64/installsst.txt. 
My SCSI tape ID was 4 and the library SCSI ID was 5. So I edited the /usr/kernel/drv/sst.conf and left the file as it was (it was already setup for a library on SCSI ID 5). 
I did not find any reference to the L280 in the /kernel/drv/st.conf so I just added the following line in the tape-config-list section : 
"SUN L280", "SUN L280 Library", "DLT-data"; 
Follow the rest of the procedure, it is straightforward. 
Unzip and untar the mtx software and follow the instructions to install it (./configure, make, make install). 
Halt your server, insert the host adapter and connect your cables. 
power on the E450. At the OK prompt, try scsi-probe-all. You should be able to see both the tape drive and the library. 

Finishing the installation and testing
Boot your server : boot -r 
Once the system is up and running, look through your /var/adm/messages file. You should see the string "changer found". If so do the following command to get your device name : 
# ls -l /dev | grep sst 
lrwxrwxrwx 1 root other 48 Nov 13 14:53 rsst5 -> ../devices/pci@4,4000/scsi@3,1/sst@5,0:character 
# 
You can now try to issue mtx commands. IMPORTANT : make sure to always specify the mtx option nobarcode. 
Here are few exemples of mtx commands with the L280 library : 
We begin with issuing an inventory check : 
# mtx -f /dev/rsst5 nobarcode inventory 
Then (and it does not have to be after the inventory check), we ask a status of the library : 
# mtx -f /dev/rsst5 nobarcode status 
Storage Changer /dev/rsst5:1 Drives, 8 Slots ( 0 Import/Export ) 
Data Transfer Element 0:Empty 
Storage Element 1:Empty 
Storage Element 2:Empty 
Storage Element 3:Empty 
Storage Element 4:Empty 
Storage Element 5:Empty 
Storage Element 6:Empty 
Storage Element 7:Full 
Storage Element 8:Full 
# 
We load slot 8 tape into the tape drive 
# mtx -f /dev/rsst5 nobarcode load 8 
The status has changed : 
# mtx -f /dev/rsst5 nobarcode status 
Storage Changer /dev/rsst5:1 Drives, 8 Slots ( 0 Import/Export ) 
Data Transfer Element 0:Full (Storage Element 8 Loaded) 
Storage Element 1:Empty 
Storage Element 2:Empty 
Storage Element 3:Empty 
Storage Element 4:Empty 
Storage Element 5:Empty 
Storage Element 6:Empty 
Storage Element 7:Full 
Storage Element 8:Empty 
# 
We Unload the tape : 
# mtx -f /dev/rsst5 nobarcode unload 
Unloading Data Transfer Element into Storage Element 8...done 
# 
NOTE : I was not able to unload the tape to a slot of my choice. It seems that this function does not work on my library. So I can only unload a tape and put it back in its original slot. 
Anyway, That's great. Congratulations to everybody involved in the development of the sst and the mtx software. Good job ! 
The only thing left is making your own scripts using these commands, or use Amanda.

Changing expirery date of a gpg key

# gpg --list-keys
/root/.gnupg/pubring.gpg
------------------------
pub 1024D/0638309A 2005-06-29 Groupe Unix (no comment) <xxxx@xxxx.com>
sub 2048g/A50FAD76 2005-06-29 [expires: 2006-06-29]

# gpg --edit-key "Groupe Unix"
gpg (GnuPG) 1.2.1; Copyright (C) 2002 Free Software Foundation, Inc.
This program comes with ABSOLUTELY NO WARRANTY.
This is free software, and you are welcome to redistribute it
under certain conditions. See the file COPYING for details.

Secret key is available.

pub 1024D/0638309A created: 2005-06-29 expires: 2006-06-29 trust: -/e
sub 2048g/A50FAD76 created: 2005-06-29 expires: 2006-06-29
(1). Groupe Unix (no comment) <xxxx@xxxx.com>

Command> expire
Changing expiration time for the primary key.
Please specify how long the key should be valid.
0 = key does not expire
<n> = key expires in n days
<n>w = key expires in n weeks
<n>m = key expires in n months
<n>y = key expires in n years
Key is valid for? (0)
Key does not expire at all
Is this correct (y/n)? y

You need a passphrase to unlock the secret key for
user: "Groupe Unix (no comment) <xxxx@xxxx.com>"
1024-bit DSA key, ID 0638309A, created 2005-06-29

pub 1024D/0638309A created: 2005-06-29 expires: never trust: -/-
sub 2048g/A50FAD76 created: 2005-06-29 expires: 2006-06-29
(1). Groupe Unix (no comment) <xxxx@xxxx.com>

Command> key 1

pub 1024D/0638309A created: 2005-06-29 expires: never trust: -/-
sub* 2048g/A50FAD76 created: 2005-06-29 expires: 2006-06-29
(1). Groupe Unix (no comment) <xxxx@xxxx.com>

Command> expire
Changing expiration time for a secondary key.
Please specify how long the key should be valid.
0 = key does not expire
<n> = key expires in n days
<n>w = key expires in n weeks
<n>m = key expires in n months
<n>y = key expires in n years
Key is valid for? (0)
Key does not expire at all
Is this correct (y/n)? y

You need a passphrase to unlock the secret key for
user: "Groupe Unix (no comment) <xxxx@xxxx.com>"
1024-bit DSA key, ID 0638309A, created 2005-06-29

pub 1024D/0638309A created: 2005-06-29 expires: never trust: -/-
sub* 2048g/A50FAD76 created: 2005-06-29 expires: never
(1). Groupe Unix (no comment) <xxxx@xxxx.com>

Command> quit
Save changes? yes
root@spco1dpt1:/root/.gnupg
# gpg --list-keys
/root/.gnupg/pubring.gpg
------------------------
pub 1024D/0638309A 2005-06-29 Groupe Unix (no comment) <xxxx@xxxx.com>
sub 2048g/A50FAD76 2005-06-29

Converting video files with ffmpeg

This is a quick reference for converting video files with ffmpeg :

From any format to dvd :
ffmpeg -i <source_file> -target dvd <target-file>

VOB to Divx using ffmpeg

I was so fed up with my children DVD being damaged after few weeks, that I decided to convert them to divX and store them on my computer. So, when the DVD is no longer playable, I play them from the computer.
I also would like to setup a nice TV interface is the future, to select them from my living roon.

Anyway, to convert a DVD to DivX, you can use the DVD O'matic software or follow the procedure below :
1 ) extract the contents to you hard disk
2 ) use the following command line to do a 2 pass encoding :
date;ffmpeg -i filename.VOB -pass 1 -passlogfile vts.log -qscale 2 -vcodec xvid VTS_02_temp.avi;date;ffmpeg -i filename.VOB -f avi -vcodec xvid -pass 2 -passlogfile vts.log -b 800 -g 300 -bf 2 -acodec mp3 -ab 128 output_filename.avi;date
or for a single pass encoding :
ffmpeg -i filename.VOB -f avi -vcodec xvid -b 800 -g 300 -bf 2 -acodec mp3 -ab 128 ouput_filename.avi

NOTE : If your video sequence is split in 2 VOB files, you can do a simple >> to create a single VOB file :
cp VTS_01_1.VOB filename.VOB
cat VTS_01_2.VOB >> filename.VOB

RPM commands

RPM is a great packager, but it might be difficult to interogate it. So here is a summary of the command I used most :

Find out which package contains a file :
$ rpm -qf /usr/bin/vim
vim-enhanced-6.3.086-0.fc4

List files in an installed package :
$ rpm -ql vim-enhanced-6.3.086-0.fc4
/etc/profile.d/vim.csh
/etc/profile.d/vim.sh
/usr/bin/ex
/usr/bin/rvim
/usr/bin/vim
/usr/bin/vimdiff
/usr/bin/vimtutor
/usr/share/man/man1/rvim.1.gz
/usr/share/man/man1/vimdiff.1.gz
/usr/share/man/man1/vimtutor.1.gz

Find which files a package contains :
$ ls vim-enh*
vim-enhanced-6.3.071-3.i386.rpm
$ rpm -qpl vim-enhanced-6.3.071-3.i386.rpm
/etc/profile.d/vim.csh
/etc/profile.d/vim.sh
/usr/bin/ex
/usr/bin/rvim
/usr/bin/vim
/usr/bin/vimdiff
/usr/bin/vimtutor
/usr/share/man/man1/rvim.1.gz
/usr/share/man/man1/vimdiff.1.gz
/usr/share/man/man1/vimtutor.1.gz

Multimedia keyboard on Fedora Core 4

Here is how to setup your multimedia keyboard if your are using Fedora Core 4 and KDE.

1 ) Get the keycode of your multimedia keys

Start xev. This program traps the keyboard events and display information about them. Among this information you will find the keycode which will be needed later.
A sample of the output is :
KeyRelease event, serial 28, synthetic NO, window 0x1200001,
root 0x48, subw 0x0, time 1985516936, (163,-13), root:(166,60),
state 0x10, keycode 162 (keysym 0x1008ff14, XF86AudioPlay), same_screen YES,
XLookupString gives 0 bytes:
...

KeyRelease event, serial 31, synthetic NO, window 0x1200001,
root 0x48, subw 0x0, time 1985521875, (163,-13), root:(166,60),
state 0x10, keycode 234 (keysym 0x1008ff16, XF86AudioPrev), same_screen YES,
XLookupString gives 0 bytes:
...

KeyRelease event, serial 31, synthetic NO, window 0x1200001,
root 0x48, subw 0x0, time 1985522575, (163,-13), root:(166,60),
state 0x10, keycode 233 (keysym 0x1008ff17, XF86AudioNext), same_screen YES,
XLookupString gives 0 bytes:

Here I pressed the Play/Pause key, then next and previous key on my keyboard.
From the output we see that Play/Pause Key has a keycode of 162.

Do the same for the +, - and mute buttons.

2 ) Creating the .Xmodmap file
Create the file and populate it with the keycode found previously :
keycode 234 = XF86AudioPrev
keycode 162 = XF86AudioPlay
keycode 233 = XF86AudioNext
keycode 174 = XF86AudioLowerVolume
keycode 176 = XF86AudioRaiseVolume
keycode 160 = XF86AudioMute

For other XF86 references, look <a href="http://wiki.linuxquestions.org/wiki/XF86_keyboard_symbols">here</a>.

Once this file is created, load it : xmodmap ~/.Xmodmap and don't forget to add to your rc files so it will be loaded automatically when you loggon.

3 ) Configure keyboard in Control Center
We have 2 categories of command to define here. The first ones will concern the player. We will do it for Amarok but it works with other players (xmms, noatun...). The second ones will control the volume. This will be effective system wide.

3.1 ) Controling the Player
Open Control Center and click on "Regional and Accessibility", then "Input Actions"
Create a new group ( I called mine Multimedia keys ). Then create a new action.
Give it a name and select Keyboard shortcut -> DCOP call (simple) as the action type.
Click on keyboard shortcut, then map the key you want by pressing it.
Finally click on DCOP call Settings. Remote application will be amarok, Remote object will be player, Call function will be playPause. If you don't know these variables, you can click on "run KDCOP" and browse remote application, object and function.
Map as many multimedia keys as you wish then click apply.

<img border='0' hspace='5' src='/blog/uploads/xmodmap_khotkeys.jpg' alt='' />

3.2 ) Controlling the sound
The sound is controled via the amixer command. It is extremely simple. To mute, you can type amixer set Master toggle. To lower volume, type amixer set Master 1- and to raise volume, amixer set Master 1+.
So what we need to do is to add these keys/actions in khotkeys.
Go back to Control Center and add a new action in "Input Actions".
Give it a name and select "Keyboard shortcut -> Command/URL (simple)"
Map your key in Keyboard shortcut
Then type the command in "Command/URL to execute".

<img width='950' height='531' border='0' hspace='5' src='/blog/uploads/xmodmap_khotkeys2.jpg' alt='' />

Hardening your linux box

This is a tipical configuration of iptables :

#!/bin/ksh

# On re-initialise les tables
iptables -F

# On ouvre le port pour CUPS seulement pour les requetes locales
iptables -A INPUT -s localhost -p tcp --destination-port 631 -j ACCEPT
# On autorise les stations des collegues
iptables -A INPUT -s xx.xx.xx.xx -p tcp --destination-port 22 -j ACCEPT
iptables -A INPUT -s xx.xx.xx.xx -p tcp --destination-port 22 -j ACCEPT

# On autorise la connection au port VNC/RemoteDesktop 3389
iptables -A INPUT -p tcp --destination-port 3389 -j ACCEPT

# On drop tout le reste
iptables -A INPUT -p tcp --syn -j DROP

# On liste les tables
iptables -L
#!/bin/ksh

# On re-initialise les tables
iptables -F

# On ouvre le port pour CUPS seulement pour les requetes locales
iptables -A INPUT -s localhost -p tcp --destination-port 631 -j ACCEPT
# On autorise les stations des collegues
iptables -A INPUT -s 10.160.100.40 -p tcp --destination-port 22 -j ACCEPT
iptables -A INPUT -s 10.160.100.31 -p tcp --destination-port 22 -j ACCEPT

# On autorise la connection au port VNC/RemoteDesktop 3389
iptables -A INPUT -p tcp --destination-port 3389 -j ACCEPT

# On drop tout le reste
iptables -A INPUT -p tcp --syn -j DROP

# On liste les tables
iptables -L

Running Windows XP on Fedora Core 4 with qemu

Installing QEMU with Accelerator Module on Fedora Core 4

Once again, this is a doc I found on the internet. You can see the original page at <a href="http://www.brandonhutchinson.com/Installing_QEMU_with_Accelerator_Module_on_Fedora_Core.html">this link</a>.

The following are step-by-step instructions for installing QEMU with the Accelerator Module (kqemu) on Fedora Core 4. I also provide instructions for installing a Windows XP Professional guest system for use with QEMU.

1. Download and extract QEMU.
wget http://fabrice.bellard.free.fr/qemu/qemu-0.7.0.tar.gz
tar zxvf qemu-0.7.0.tar.gz

2. Download and extract the Accelerator Module in the QEMU directory.
cd qemu-0.7.0
wget http://fabrice.bellard.free.fr/qemu/kqemu-0.6.2-1.tar.gz
tar zxvf kqemu-0.6.2-1.tar.gz

3. Configure QEMU to build with Accelerator Module support using gcc 3.2.3. QEMU 0.7.0 does not compile with Fedora Core 4's gcc 4.0.0. If you do not have gcc32, install it with the yum -y install compat-gcc-32 as root.

In this example, I will install QEMU in /usr/local/qemu instead of its default location.
./configure --prefix=/usr/local/qemu --cc=gcc32

If you see:
SDL support no
SDL static link no

you must install SDL-devel and its dependencies with the yum -y install SDL-devel command as root.

4. Build QEMU.
make

5. Install QEMU.
make install (as root)

Installing a Windows XP Professional guest system
In this example, I have a bootable Windows XP Professional CD in /dev/cdrom, the logical path for my CD-ROM drive.

1. Create a disk image for the guest system. I'll create a 10 GB disk image named hd.img in my home directory for the guest system.
/usr/local/qemu/bin/qemu-img create hd.img 10G

2. Install the guest system from the bootable Windows XP Professional CD.
/usr/local/qemu/bin/qemu -boot d -hda ~/hd.img -localtime

3. After the system is installed, turn off the guest system, and boot the guest system from the disk image. I use the -user-net parameter to enable userspace networking.
/usr/local/qemu/bin/qemu -boot c -hda ~/hd.img -localtime -user-net

4. Creating a SAMBA user :
smbpasswd -a jboismar

Replacing a disk in a mirrored rootvg

In the following example, an RS6000 has 3 disks, 2 of which have the AIX
filesystems mirrored on. The boolist contains both hdisk0 and hdisk1.
There are no other logical volumes in rootvg other than the AIX system
logical volumes. hdisk0 has failed and need replacing, both hdisk0 and hdisk1
are in "Hot Swap" carriers and therefore the machine does not need shutting
down.

lspv

hdisk0 00522d5f22e3b29d rootvg
hdisk1 00522d5f90e66fd2 rootvg
hdisk2 00522df586d454c3 datavg

lsvg -l rootvg

rootvg:
LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
hd6 paging 4 8 2 open/syncd N/A
hd5 boot 1 2 2 closed/syncd N/A
hd8 jfslog 1 2 2 open/syncd N/A
hd4 jfs 1 2 2 open/syncd /
hd2 jfs 12 24 2 open/syncd /usr
hd9var jfs 1 2 2 open/syncd /var
hd3 jfs 2 4 2 open/syncd /tmp
hd1 jfs 1 2 2 open/syncd /home

1, Reduce the logical volume copies from both disks to hdisk1 only :-

rmlvcopy hd6 1 hdisk0
rmlvcopy hd5 1 hdisk0
rmlvcopy hd8 1 hdisk0
rmlvcopy hd4 1 hdisk0
rmlvcopy hd2 1 hdisk0
rmlvcopy hd9var 1 hdisk0
rmlvcopy hd3 1 hdisk0
rmlvcopy hd1 1 hdisk0

2, Check that no logical volumes are left on hdisk0 :-

lspv -p hdisk0

hdisk0:
PP RANGE STATE REGION LV ID TYPE MOUNT POINT
1-101 free outer edge
102-201 free outer middle
202-301 free center
302-401 free inner middle
402-501 free inner edge

3, Remove the volume group from hdisk0

reducevg -df rootvg hdisk0

4, Recreate the boot logical volume on hdisk1, and reset bootlist:-

bosboot -a -d /dev/hdisk1
bootlist -m normal rmt0 cd0 hdisk1

5, Check that everything has been removed from hdisk0 :-

lspv

hdisk0 00522d5f22e3b29d None
hdisk1 00522d5f90e66fd2 rootvg
hdisk2 00522df586d454c3 datavg

6, Delete hdisk0 :-

rmdev -l hdisk0 -d

7, Remove the failed hard drive and replace with a new hard drive.

8, Configure the new disk drive :-

cfgmgr

9, Check new hard drive is present :-

lspv

10, Include the new hdisk in root volume group :-

extendvg rootvg hdisk? (where hdisk? is the new hard disk)

11, Re-create the mirror :-

mirrorvg rootvg hdisk? (where hdisk? is the new hard disk)

12, Syncronise the mirror :-

syncvg -v rootvg

13, Reset the bootlist :-

bootlist -m normal rmt0 cd0 hdisk0 hdisk1

14, Turn off Quorum checking on rootvg :-

chvg -Q n rootvg

Enabling command history search

On Fedora Core 4, the command history search function is not bind to any key. Therefore if you want to find a command you typed in, you will have to go through them all.

A solution is to create a binding for this search function. Edit the file /etc/inputrc and add the following 2 lines :
"\C-f": history-search-backward
"\C-g": history-search-forward

Logout and log in again and Voila !
Type the beginning of the command you want to find and press CTRL-f ou CRTL-g.

Finding out how long ago a daemon wrote to a log file

In my day to day job, I often have to write monitoring scripts. Recently we had to write a script that would check if a daemon wrote in its log in the last 5 minutes. This particular daemon had the habit of freezing. So you would see it in the process list but it would not do anything.
Therefore, to ensure that this process is alive and working, we have to check the last entry it creates in the log file and see if it is newer than 5 mns ( It is supposed to write at least 4 entries per minutes). Here is the litlle perl script to do that check :

##########################################
# Nouveau check pour DaemonA
&debug ( "=-=-=-=-= Verification de DaemonA =-=-=-=-= " );
$last_Order=`grep "DaemonA" /var/logs/efixeng/enginelog4j.log | tail -1`;
if ( $last_Order == "" )
{
&debug ( "We did not find any log entry for DaemonA");
$ORDER_TIME = 0;
} else {
&debug ( "last log entry was $last_Order" );
($order_date,$order_heure,$trash) = split (/ /,$last_Order);
&debug ( "date is $order_date, heure is $order_heure");
# Setting up date
($ohours,$ominutes,$osecondes)=split(/:/,$order_heure);
($oyear,$omonth,$oday)=split(/\./,$order_date);
$ORDER_TIME = timelocal($osecondes, $ominutes, $ohours, $oday, $omonth, $oyear);
&debug ( "Epoch time for the last occurence of DaemonA is $ORDER_TIME");
}

$CURRENT_TIME = timelocal(`date +%S`,`date +%M`, `date +%H`, `date +%d`, `date +%m`, `date +%Y`);
&debug ( "Current Epoch time is $CURRENT_TIME");

$diff = $CURRENT_TIME - $ORDER_TIME;
if ( $diff < 300 )
{
&debug ( "Last Occurence of DaemonA was $diff secondes ago. Everything is OK");
$message .= "\nLe Daemon DaemonA a ecrit une entree il y a $diff secondes. tout est OK.\n";
} else {
&debug ( "[ERREUR!] Last Occurence of DaemonA was $diff secondes ago.");
$summary .= "\n[ERREUR!] Le DaemonA Daemon ne roule pas ($cdate $cheure) !!";
$summary .= "\n[ERREUR!] La derniere ecriture dans le log etait il y a $diff secondes !!";
$color = "red";
};
# fin de la verification
##########################################

nic speed on a solaris

The following script will give you a Sun Solaris network interface information like link speed...

#!/bin/sh

# Only the root user can run the ndd commands
if [ "`/usr/bin/id | /usr/bin/cut -c1-5`" != "uid=0" ] ; then
echo "You must be the root user to run `basename $0`."
exit 1
fi

# Print column header information
/usr/bin/echo "Interface\tSpeed\t\tDuplex\t\tAutoneg"
/usr/bin/echo "---------\t-----\t\t------\t\t-------"

# Determine the speed and duplex for each live NIC on the system
for INTERFACE in `/usr/bin/netstat -i | /usr/bin/egrep -v "^Name|^lo0" \
| /usr/bin/awk '{print $1}' | /usr/bin/sort | /usr/bin/uniq`
do
# Special handling for "ce" interfaces
if [ "`/usr/bin/echo $INTERFACE \
| /usr/bin/awk '/^ce[0-9]+/ { print }'`" ] ; then
# Determine the ce interface number
INSTANCE=`/usr/bin/echo $INTERFACE | cut -c 3-`
DUPLEX=`/usr/bin/kstat ce:$INSTANCE | /usr/bin/grep link_duplex \
| /usr/bin/awk '{ print $2 }'`
case "$DUPLEX" in
1) DUPLEX="half" ;;
2) DUPLEX="full" ;;
esac
SPEED=`/usr/bin/kstat ce:$INSTANCE | /usr/bin/grep link_speed \
| /usr/bin/awk '{ print $2 }'`
case "$SPEED" in
10) SPEED="10 Mbit/s" ;;
100) SPEED="100 Mbit/s" ;;
1000) SPEED="1 Gbit/s" ;;
esac
AUTONEG=`/usr/bin/kstat ce:$INSTANCE | /usr/bin/grep adv_cap_autoneg \
| /usr/bin/awk '{ print $2 }'`
case "$AUTONEG" in
0) AUTONEG="NO" ;;
1) AUTONEG="YES" ;;
esac

# Special handling for "bge" interfaces
elif [ "`/usr/bin/echo $INTERFACE \
| /usr/bin/awk '/^bge[0-9]+/ { print }'`" ] ; then
BGE_INT_LINE_NO=`/usr/bin/kstat bge | /usr/bin/grep -n $INTERFACE \
| /usr/bin/awk -F: '{print $1}'`
BGE_INT_DUPLEX_LINE_NO=`/usr/bin/expr $BGE_INT_LINE_NO + 9`
BGE_INT_SPEED_LINE_NO=`/usr/bin/expr $BGE_INT_LINE_NO + 14`
DUPLEX=`/usr/bin/kstat bge | /usr/bin/awk 'NR == LINE { print $2 }' \
LINE=$BGE_INT_DUPLEX_LINE_NO`
SPEED=`/usr/bin/kstat bge | /usr/bin/awk 'NR == LINE { print $2 }' \
LINE=$BGE_INT_SPEED_LINE_NO`
case "$SPEED" in
10000000) SPEED="10 Mbit/s" ;;
100000000) SPEED="100 Mbit/s" ;;
1000000000) SPEED="1 Gbit/s" ;;
esac
# All other interfaces
else
INTERFACE_TYPE=`/usr/bin/echo $INTERFACE | /usr/bin/sed -e "s/[0-9]*$//"`
INSTANCE=`/usr/bin/echo $INTERFACE | /usr/bin/sed -e "s/^[a-z]*//"`
/usr/sbin/ndd -set /dev/$INTERFACE_TYPE instance $INSTANCE
SPEED=`/usr/sbin/ndd -get /dev/$INTERFACE_TYPE link_speed`
case "$SPEED" in
0) SPEED="10 Mbit/s" ;;
1) SPEED="100 Mbit/s" ;;
1000) SPEED="1 Gbit/s" ;;
esac
DUPLEX=`/usr/sbin/ndd -get /dev/$INTERFACE_TYPE link_mode`
case "$DUPLEX" in
0) DUPLEX="half" ;;
1) DUPLEX="full" ;;
*) DUPLEX="" ;;
esac
AUTONEG=`/usr/sbin/ndd -get /dev/$INTERFACE_TYPE adv_autoneg_cap`
case "$AUTONEG" in
0) AUTONEG="NO" ;;
1) AUTONEG="YES" ;;
esac
fi
/usr/bin/echo "$INTERFACE\t\t$SPEED\t$DUPLEX\t\t$AUTONEG"
done

Looking for information about a stale partition

When you have mirrors for your data, it might be a little difficult to find out which disk has the stale partition. The following few commands can help you find out quickly where the stale partition is located physically.

1 ) List of volumes groups
root@myhost:/root# lsvg
rootvg
datavg

2 ) Which volume group has a stale partition ?
root@myhost:/root# lsvg -l rootvg | grep stale
hd2 jfs 108 216 2 open/stale /usr
root@myhost:/root# lsvg -l datavg | grep stale

3 ) Which physical volumes are defined in that VG ?
root@myhost:/root# lsvg -p rootvg
rootvg:
PV_NAME PV STATE TOTAL PPs FREE PPs FREE DISTRIBUTION
hdisk0 active 542 81 00..00..00..00..81
hdisk1 active 542 97 08..02..00..00..87

4 ) Which volume group has the stale partition ?
root@myhost:/root# lspv -p hdisk1 | grep stale
300-300 stale center hd2 jfs /usr
root@myhost:/root# lspv -p hdisk0 | grep stale

Other commandes you might find usefull :
root@myhost:/root# lsvg -l rootvg
rootvg:
LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
hd5 boot 1 2 2 closed/syncd N/A
hd6 paging 32 64 2 open/syncd N/A
hd8 jfslog 1 2 2 open/syncd N/A
hd4 jfs 1 2 2 open/syncd /
hd2 jfs 108 216 2 open/stale /usr
hd9var jfs 128 256 2 open/syncd /var
hd3 jfs 8 16 2 open/syncd /tmp
hd1 jfs 10 20 2 open/syncd /home
hd7 dump 16 16 1 open/syncd N/A

To see a mapping of this logical volume on disk :
root@myhost:/root# lslv -m hd2
hd2:/usr
LP PP1 PV1 PP2 PV2 PP3 PV3
0001 0220 hdisk0 0220 hdisk1
0002 0221 hdisk0 0221 hdisk1
0003 0222 hdisk0 0222 hdisk1
0004 0223 hdisk0 0223 hdisk1
0005 0224 hdisk0 0224 hdisk1
0006 0225 hdisk0 0225 hdisk1
0007 0226 hdisk0 0226 hdisk1
0008 0227 hdisk0 0227 hdisk1
0009 0228 hdisk0 0228 hdisk1
...

To see which part of the mirror has the stale partition :
root@myhost:/root# lslv -p hdisk0
hdisk0:::
USED USED USED USED USED USED USED USED USED USED 1-10
USED USED USED USED USED USED USED USED USED USED 11-20
USED USED USED USED USED USED USED USED USED USED 21-30
USED USED USED USED USED USED USED USED USED USED 31-40
USED USED USED USED USED USED USED USED USED USED 41-50
USED USED USED USED USED USED USED USED USED USED 51-60
USED USED USED USED USED USED USED USED USED USED 61-70
...

root@myhost:/root# lslv -p hdisk1
hdisk1:::
USED FREE FREE FREE FREE FREE FREE FREE FREE USED 1-10
USED USED USED USED USED USED USED USED USED USED 11-20
...
USED USED USED USED USED USED USED USED USED USED 278-287
USED USED USED USED USED USED USED USED USED USED 288-297
USED USED STALE USED USED USED USED USED USED USED 298-307
USED USED USED USED USED USED USED USED USED USED 308-317
...
So the stale partition is on hdisk1.

Another command to be user is :
lspv -p hdisk1
hdisk1:
PP RANGE STATE REGION LV NAME TYPE MOUNT POINT
1-1 used outer edge hd5 boot N/A
2-9 free outer edge
10-15 used outer edge lvlogsihs4 jfs /var/logs/ihs4
16-45 used outer edge lvarchive jfs /var/archive
46-52 used outer edge lvdomidata2 jfs /usr/domino/data2
53-106 used outer edge paging00 paging N/A
107-109 used outer edge lvdomilogs jfs /var/logs/domino
110-141 used outer middle hd6 paging N/A
142-145 used outer middle lvwaslogs jfs /var/logs/was
146-147 used outer middle lvhttplogs jfs /var/logs/ihs
148-187 used outer middle lvdomidata jfs /usr/domino/data1
188-197 used outer middle lvdomibin jfs /usr/domino/bin
198-199 free outer middle
200-209 used outer middle hd1 jfs /home
210-217 used outer middle hd3 jfs /tmp
218-218 used center hd8 jfslog N/A
219-219 used center hd4 jfs /
220-299 used center hd2 jfs /usr
300-300 stale center hd2 jfs /usr
301-325 used center hd2 jfs /usr
326-327 used inner middle hd2 jfs /usr
328-433 used inner middle hd9var jfs /var
434-455 used inner edge hd9var jfs /var
456-542 free inner edge

Image borders in HTML

There seems to be a bug in Firefox when we try to display images side by side. I wrote a PHP application to show digital pictures to friends and family. I use a group of images setup side by side to simulate a picture frame around my digital picture :
<center><img width='230' height='201' border='0' hspace='5' src='/blog/uploads/meimei_borber_ok.jpg' alt='' /></center>

Firefox does not want to display 2 images side by side without space. This is how my page is displayed with Firefox :

<center><img width='232' height='239' border='0' hspace='5' src='/blog/uploads/meimei_border_nonok.jpg' alt='' /></center>

After looking everywhere on the web, I resigned myself of changing the display. If the browser is not Internet Explorer, I do not display these nice picture frames, I display simple image borders. I found a very good page about setting up borders in CSS : <a href="http://www.mandarindesign.com/boxes.html">click here</a>.

So until I found a fix, I'll be playing with image borders in CSS.

Nicer fonts with Fedora 4

This is a reproduction of a page I gound on the internet. Here is the <a href="http://cri.ch/linux/docs/sk0017.html">original page</a>.

Nicer fonts for Fedora Core 4

Author: Sven Knispel
Updated: 01-11-2005
Feedback welcome: linux@cri.ch
Free service provided by: www.cri.ch

Ever since I am using Fedora Core I have been complaining about the poor rendering of fonts: they look ugly and unsharp.
One week ago I switched my desktop PC from Windows to Linux and I really couldn't stand it so I started a little research.

The reason of the poor rendering is that FreeType is compiled by default with the bytecode interpreter switched off. I didn't find the real reason for that but it seems related to some patent issues.
Fortunately it is quite simple to turn the bytecode interpreter on by recompiling freetype after a slight change.
This article is about recompiling freetype with the bytecode interpreter switched on:

* get the source rpm
* make the required changed
* recompile and install the modified freetype version

1. Getting the sources

First we must heck for current version of freetype:
# rpm --query freetype
# freetype-2.1.9-2
Then download the sources, e.g. from rpm.bone: http://rpm.pbone.net/index.php3/stat/4/idpl/1981137/com/freetype-2.1.9-2.i386.rpm.html
(I suggest downloading the binary rpm as well in case of dammage as we will freshen the installed one)
Install the sources:
sudo rpm -ivh ./Desktop/freetype-2.1.9-2.src.rpm
2. Enabling the bytecode interpreter

Change to the directory /usr/src/redhat/SPECS and edit freetype.spec.
What we are looking for is: %define without_bytecode_interpreter 1 and we want to replace it by: %define without_bytecode_interpreter 0 in order to enable the bytecode interpreter.
3. Rebuild and install

After having changed the code we need to rebuild the RPMs:
rpmbuild -bb freetype.spec (if you get errors make sure that XFree86-devel is installed as freetype depends on it)
If the build was successfull the result can be found in /usr/src/redhat/RPMS/i386/: [~]$ cd /usr/src/redhat/RPMS/i386/
[i386]$ ls -l
total 3268
-rw-r--r-- 1 root root 824001 Oct 21 15:28 freetype-2.1.9-2.i386.rpm
-rw-r--r-- 1 root root 1781435 Oct 21 15:28 freetype-debuginfo-2.1.9-2.i386.rpm
-rw-r--r-- 1 root root 105464 Oct 21 15:28 freetype-demos-2.1.9-2.i386.rpm
-rw-r--r-- 1 root root 585133 Oct 21 15:28 freetype-devel-2.1.9-2.i386.rpm
-rw-r--r-- 1 root root 25476 Oct 21 15:28 freetype-utils-2.1.9-2.i386.rpm
Now check for the ones you need to reinstall with rpm --query <package-name> (e.g. rpm --query freetype-devel) and reinstall whatever is required with the force option as they are already installed in the original version: rpm -Uvh --force freetype-2.1.9-2.i386.rpm
rpm -Uvh --force freetype-devel-2.1.9-2.i386.rpm
That's it!
Now just restart X and enjoy...
4. Settings

If you use a TFT I recommend deactivating anti-aliasing for even crisper fonts.
4. Results

And here is the comparison before and after:
before after
5. Consistent font for all applications (updated 31.10.2005)

If you are using KDE you may have noticed that the settings from KDE are not inherited to Qt applications. Especially Firefox kept showing bigger fonts not corresponding to me KDE font-settings.

After some research I found out how to force the fonts for non KDE applications:

* the gnome font-settings must be changed. This can be done by exectuting gnome-font-properties and select e.g. Tahoma 8 as default font
* under KDE the gnome-settings-daemon mus be started: this can be done e.g. by adding a symlink to /usr/libexec/gnome-settings-daemon from ~/.kde/Autostart (ln -s /usr/libexec/gnome-settings-daemon ~/.kde/Autostart/gnome-settings-daemon)

After these changes all my apps have neat and consistent fonts...

Capturing DV video from camcoder

After spending few weeks in Taiwan at my family in law's place, I had to transfer some videos I took with my Bro's miniDV video camera. I found very usefull information on how to set up Linux <a href="https://www.redhat.com/archives/fedora-list/2004-December/msg06925.html">here</a>. To make sure these information won't go away, I extracted what I have used from this link and present them below :

As root:
$ cd /dev/
$ ./MAKEDEV raw1394
$ mknod -m 666 /dev/dv1394 c 171 32
$ chmod 666 raw1394

Then add DAG to your yum repository list:
$ cat /etc/yum.repos.d/dag.repo
[dag]
name=Dag RPM Repository for Fedora Core
baseurl=http://apt.sw.be/fedora/$releasever/en/$basearch/dag

Then install kino and associated tools:
$ yum install kino
$ yum install dvgrab

Then plug in the card and the cam camcorder, set the camera to the
"play" mode, and load the necessary modules:
$ modprobe ieee1394
$ modprobe ohci1394
$ modprobe raw1394 dv1394

Finally, as a normal user:
$ dvgrab file001

and you should start getting file transfers from the camcorder.

More info is available at the following places:

http://kino.schirmacher.de/
http://www.linux1394.org/
http://www.syba.com/us/en/product/43/05/04/index.html
http://www.linux1394.org/hcl.php?class_id=3
http://www.linux1394.org/hcl.php?class_id=1

Exec statement in KDE

A very nice feature of KDE environment is to personalized servicemenu (right click in Knoqueror). You can define an action to do but I was looking for arguments to pass by to the command. I found them <a href="http://standards.freedesktop.org/desktop-entry-spec/desktop-entry-spec-0.9.4.html">here</a>

Recognized fields are as follows:
%f A single file name, even if multiple files are selected. The system reading the desktop entry should recognize that the program in question cannot handle multiple file arguments, and it should should probably spawn and execute multiple copies of a program for each selected file if the program is not able to handle additional file arguments. If files are not on the local file system (i.e. are on HTTP or FTP locations), the files will be copied to the local file system and %f will be expanded to point at the temporary file. Used for programs that do not understand the URL syntax.
%F A list of files. Use for apps that can open several local files at once.
%u A single URL.
%U A list of URLs.
%d Directory containing the file that would be passed in a %f field.
%D List of directories containing the files that would be passed in to a %F field.
%n A single filename (without path).
%N A list of filenames (without paths).
%i The Icon field of the desktop entry expanded as two parameters, first --icon and then the contents of the Icon field. Should not expand as any parameters if the Icon field is empty or missing.
%c The translated Name field associated with the desktop entry.
%k The location of the desktop file as either a URI (if for example gotten from the vfolder system) or a local filename or empty if no location is known.
%v The name of the Device entry in the desktop file.

a tree view of directories

Working in a production environment, we often need to find out which directories take up all space. At first, I was using the command df -k . | sort -u but the output is not very easy to look at. I decided to create a small perl script that would display the size of all the directories in the current location. Here is an exemple :

<img width='450' height='379' border='0' hspace='5' src='/blog/uploads/jbdu01.gif' alt='' />

You can also call the script with a depth argument. If you want to have the size of 2 subdir levels (current level and one sub level), call the script with 2 :

<img width='499' height='464' border='0' hspace='5' src='/blog/uploads/jbdu02.gif' alt='' />

This is a very simple script but it helps me a lot. That's why I decided to publish it on my blog. Use it as you wish !!! It works as is on Linux (Fedora core 3 and 4), AIX 4.X and 5.X, and Solaris 8 and 9.

<a href="/blog/uploads/jbdu" title="jbdu" target="_blank">You can download the code from here.

How to run Lotus Notes with Wine

1. Download and install the 20041019 versions of the wine, libwine, and wine-utils binaries from this Sourceforge page (newer versions may have regression bug #2660).

2. Download and install the latest version of winesetuptk using apt-get or Synaptic.

3. Run winesetup (as you, not root) and accept the defaults.

4. Copy the Notes folder and all subfolders from a working copy of the Lotus Notes [6.5.1 or later] client on Windows to the equivalent directory in your wine installation (for example, if it was "c:\lotus\notes" under Windows, it should be copied to "~/.wine/c/lotus/notes" on Linux).

Make sure that your notes.ini file is in your Notes program directory (the same directory that nlnotes.exe is in). If it's not, you should find it (probably in the Windows directory) and copy it there.

Also, while you're looking at the notes.ini file, make sure that any AddinMenus= or EXTMGR_ADDINS= settings are commented out. They may not hurt anything if they're there, but you should definitely test without them first.

5. Copy mfc42.dll and msvcp60.dll from your "c:\windows\system32" directory in Windows to "~/.wine/c/windows/system" on Linux.

6. Add the following sections to your ~/.wine/config file:

[AppDefaults\\nlnotes.exe\\Version]
"Windows" = "win98"
[AppDefaults\\nlnotes.exe\\x11drv]
"Managed" = "Y"
"DesktopDoubleBuffered" = "Y"

If there are any other sections that were already there for [AppDefaults\\nlnotes.exe], [AppDefaults\\notes.exe], or [AppDefaults\\nhldaemn.exe], just comment them out. We don't need to override any DLLs or set screen resolution or anything with this setup.

Along those lines, make sure any global "Resolution=" settings are commented out. You don't need that either.

7. Try to run Notes. From a terminal window, type:

wine "c:\lotus\notes\nlnotes.exe"

Obviously, you should use whatever path you copied the Notes installation to (c:\lotus\notes, c:\Program Files\lotus\notes, whatever).

Using ssh keys with Filezilla

One of the best FTP client I used under windows is FileZilla. It is simple but works very well. You can download it <a href="http://filezilla.sourceforge.net/">here.</a>
It is great to connect to old FTP sites and can now be used to connect to site using FTP over ssh. It is pretty straight forward if you can login using user/password authentification but it is more tricky if you want to use public/private ssh keys.
The trick is that FileZilla uses modules from putty (the best free telnet/ssh client for windows) to authenticate using keys.

Here is the complete (I hope !) configuration of FileZilla and putty to use ssh keys.

1 ) download the components
Go to <a href="http://www.chiark.greenend.org.uk/~sgtatham/putty/download.html">putty download page</a> and download either putty.zip (which contains everything needed) or download the following components :
- PuTTY : The telnet/ssh client itself
- PuTTYgen : The ssh key generator
- Pageant : The ssh agent to handle passphrase ( I will explain it later ).

2 ) Generating your keys
The first step is to create your public/private key pair that you will be using to access sites. We will only consider RSA ssh keys in this document. Open PuTTYgen. 

<img width='' height='' border='0' hspace='5' align='left' src='http://moon.homeunix.com:8080/blog/uploads/putty01.jpg' alt='' /> 

Click on "Generate". You will have to move your mouse to generate random numbers. When it 's finished you will see the following screen : 

<img width='' height='' border='0' hspace='5' align='left' src='http://moon.homeunix.com:8080/blog/uploads/putty02.jpg' alt='' /> 

You can modify the Key Comment with whatever you want. It is a good idea to put something like : MyWindowsXP key.... This description is for you, it is not parsed by the ssh system.
The next field is extremelly important : Please, USE A PASSPHRASE.
When everything is done, click on "Save public key". It is a good idea to name your file id_rsa.pub. Although this is not mandatory (you can give any name you want), this would respect the standard name for rsa key file.
Then click on save private key. You can name this file id_rsa.
A quick explanation on these keys. The 2 keys you generated work together. You always keep you private key for yourself. You will never have to send it to somebody. It is your property so keep it secure ;-) ! You public key will have to be sent to every single system you access.

The key generation process if finished. You can close PuTTYgen.
Now is the right time to add your key to the destination account you want to access. Let's say you want to access host1 and account account1. You need to add (or the host administrator) your public key to the file account1@host1:~/.ssh/authorized_keys2.

3 ) configuring PuTTY
Open PuTTY and configure the host you want to access. Click on ssh->auth and browse to get your private key file : 

<img width='' height='' border='0' hspace='5' align='left' src='http://moon.homeunix.com:8080/blog/uploads/putty04.jpg' alt='' /> 

Now click back to session, give it a name and save it.

4) Configuring Pageant
If you try to connect to your host using PuTTY and your key, the system will ask for your passphrase. Unfortunately, FileZilla does not ask for a passphrase. So we need the help of Pageant to provide it for you to FileZilla. Pageant is called an ssh-agent. It is used to store your key and your passphrase and give it to client like PuTTY and FileZilla. This is very convenient if you work with a lot of servers. You enter your key and passphrase once, and you can login to any host without typing it again. Each time you start a new session, PuTTY will get its authentification info directly from Pageant.
Let's configure Pageant. Run it and look for it in your task bar. It is the little icon of a computer with a hat on :
 

<img width='' height='' border='0' hspace='5' align='left' src='http://moon.homeunix.com:8080/blog/uploads/Pageant01.jpg' alt='' />
 

Right click on it add key. Browse to get your private key. It will then ask you for you passphrase.

5 ) Using FileZilla to connect to your site

Finally you can open FileZilla. Configure your session. You should be able to connect using FTP over ssh2 to your host.

Thursday, July 12, 2007

Sending out email through your provider with sasl authentification

Over the past years, ISPs started to require email client to authenticate before sending a mail through their mail server. This was an attempt to prevent spamming.
For that time on, no more default configuration of sendmail ! (which actually is much better ;-). Alhtough the configuration is not difficult, it took me sometimes to go through the voluminous mail documentation. Eventually I find out what to do to enable authentification and to remap my local sender email address (root@localhost.localdomain) to a more friendly address (jerome.boismartel@gmail.com).

When I installed my new server using Fedora Core 4, I had once again to figure out what to install and how to configure it. This is the steps I followed :

1 ) Finding out what is currently on the system
[root@localhost log]# rpm -qa | grep sasl
gnu-crypto-sasl-jdk1.4-2.0.1-1jpp_5fc
cyrus-sasl-devel-2.1.20-5
cyrus-sasl-md5-2.1.20-5
cyrus-sasl-plain-2.1.20-5
cyrus-sasl-2.1.20-5

Cyrus-sasl is the library that takes care of the authentification. As I discovered, it seems to be installed by default on the system (at least with my package selection).

I decided to use postfix as a replacement for sendmail. This is a much better mail system and it's easier to configure. You can find more about postfix <a href="http://www.postfix.org">here</a>.

2 ) Installing missing components 
I could not find a postfix package on my system. So I called up the handy yum ! :
yum install postfix

3 ) Configuration 
I edited the file /etc/postfix/main.cf to add the following lines :
smtp_sasl_auth_enable = yes
smtp_sasl_security_options = noanonymous
smtp_sasl_password_maps=static:<your account>:<your password>
smtp_use_tls = yes
smtp_generic_maps = hash:/etc/postfix/generic

This basically tells your system that your ISP requires authentification. On the third line, you specify your account name and your password.
The last line is not about athentification but it is about address remapping. As I said at the beginning of this text, I wanted my address to appear as if my email was sent from my main mail account : jerome.boismartel@gmail.com and not from root@localhost.localdomain.
To enable address remapping you need to add the last line. Then you add your mapping to /etc/postfix/generic. It is a simple text file that says : "If local address is root@localhost.localdomain, remap it to jerome.boismartel@gmail.com". Here is the file :
#cat /etc/postifx/generic
root@whatever jerome.boismartel@gmail.com

Now that you added your entry, you need to create a DB from that file. The postfix program will search the DB and not the text file. Issue the command :
postmap /etc/postfix/generic

Now you probably want to stop sendmail which is most probably running :
/etc/init.d/sendmail stop

It would be a good idea at that point to rename this startup script preventing it to start at the next reboot :
mv /etc/init.d/sendmail /etc/init.d/sendmail.old

You can now start your postfix mail :
/etc/init.d/postfix start

And send you first email to the world !
echo "SUBJECT:test" | sendmail <email>

This server was built using the latest release of Fedora Core 4. My previous server was running an old version of Redhat (7.2) and