All about Fedora

F15: Going back to eth0, eth1, ...

2011-10-04T14:30:00.002+01:00

Fedora has decided to change the names of the network interface from ethX to emX (see here).

If you decide to install a commercial software which needs to check its FLEXnet license (from Flexera) on a server running F15, you will need to provide the hostid of your server to the software company by running the lmhostid executable, which gives the MAC address of your server.

with network interfaces named like em0, em1 ..., you will get a null MAC address:
$ ./lmhostid
lmhostid - Copyright (c) 1989-2008 Acresso Software Inc. All Rights Reserved.
The FLEXnet host ID of this machine is "000000000000"

which is the MAC address of the loopback (lo) interface.

Therefore, you need to change the network interface name back to ethX. FLEXnet license manager only understand network interfaces starting with eth.

Add the kernel option biosdevname=0 in /etc/grub.conf:
$ su -
# GRUBBYLASTKERNEL=`/sbin/grubby --default-kernel`
# /sbin/grubby --update-kernel=${GRUBBYLASTKERNEL} --args='biosdevname=0' &>/dev/null

Edit file /etc/udev/rules.d/70-persistent-net.rules and add the 2 following lines:
SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="00:11:22:33:44:10", ATTR{dev_id}=="0x0", ATTR{type}=="1", KERNEL=="eth*", NAME="eth0"
SUBSYSTEM=="net", ACTION=="add", DRIVERS=="?*", ATTR{address}=="00:11:22:33:44:11", ATTR{dev_id}=="0x0", ATTR{type}=="1", KERNEL=="eth*", NAME="eth1"

Be sure to put your MAC addresses in the fields above.

Rename the ifcfg-* config files in /etc/sysconfig/network-scripts to match the eth0 and eth1 device names.

Then reboot and re-run lmhostid to get:
$ ./lmhostid
lmhostid - Copyright (c) 1989-2008 Acresso Software Inc. All Rights Reserved.
The FLEXnet host ID of this machine is ""001122334410 001122334411""

Many thanks to Racker Hacker (Getting back to using eth0 in Fedora 15).

F14 : How to enable USB 3.0 ports?

2011-09-30T11:26:00.001+01:00

USB 3.0 ports are not enabled by default on F14. It prevents users from being able to suspend their laptops. The module is called xhci_hcd. It can be called xhci.

Check with
$ lsmod | grep xhci
and change the following accordingly:
$ su -
# GRUBBYLASTKERNEL=`/sbin/grubby --default-kernel`
# /sbin/grubby --update-kernel=${GRUBBYLASTKERNEL} --args='xhci_hcd.enable=1' &>/dev/null
# echo "options xhci-hcd enable=1" > /etc/modprobe.d/xhci_hcd.conf
# echo "SUSPEND_MODULES=xhci_hcd" > /etc/pm/config.d/xhci_hcd

Read about it on the "Common bug in F14" on Fedora Project and this post on FedoraForum.

How to display Japanese characters in Google Chrome/Chromium?

2010-03-24T15:40:00.005+01:00

If you see the Japansese characters are display as squares in Google Chrome/Chromium like this:

I'm almost certain you have the google-droid fonts installed on your system:

google-droid-serif-fonts-20090906-5.fc12.noarch
google-droid-fonts-common-20090906-5.fc12.noarch
google-droid-sans-mono-fonts-20090906-5.fc12.noarch
google-droid-sans-fonts-20090906-5.fc12.noarch

Just remove them:

yum remove google-droid*

And you should be able to see the Japanese characters correctly like this:

That' all, Folks!

Funny! Google fonts not compatible with Google Chrome/Chromium!

Google wave: What OS are you using?

2009-12-04T23:57:00.000+01:00

Here is Google wave entitled What OS you are using?. There is a poll in that wave showing that Fedora is used by only 2% of the peolple who answered the poll. Windows users totalizes 47.1%, Linux users, 29.1% and Mac OS X users, 21.2%. And 58% of the people use a 64-bit computer.

So either you Fedora people don't have a google wave account or you didn't find the wave or ..., but those 2 lousy % seem to me ridiculous. We need to increase that number.

If you have a google wave account, please follow the link above and vote! Fedora needs you, guys.

F12: Yum new feature: history

2009-11-20T23:55:00.005+01:00

With Fedora 12 (which was released a few days ago), a new version of yum was released too (3.2.25). One of the new features of yum I always dreamed about was the possibility to undo was I installed or removed. This is done now with the 'history' command of yum eventually associated with one of the following options 'info|list|summary|redo|undo|new'.

Here are a few examples:

The following lists all actions performed with yum:

# yum history
Loaded plugins: downloadonly, fastestmirror, keys, presto, refresh-packagekit
ID | Login user | Date and time | Action(s) | Altered
-------------------------------------------------------------------------------
   28 | System    | 2009-11-20 22:55 | Install | 12
   27 | John ...   | 2009-11-20 22:51 | Erase | 12
   26 | John ...   | 2009-11-20 22:37 | Install | 12
   25 | System    | 2009-11-20 21:26 | Update | 27
   24 | System    | 2009-11-20 08:12 | Install | 1
   23 | System    | 2009-11-18 23:48 | Install | 11
   22 | System    | 2009-11-18 23:09 | Install | 27
   21 | System    | 2009-11-18 22:41 | Install | 1
   20 | System    | 2009-11-18 22:39 | Install | 2
   19 | System    | 2009-11-18 22:36 | Install | 11
   18 | John  ...   | 2009-11-18 22:31 | Install | 3
   17 | John  ...   | 2009-11-18 22:30 | Install | 7
   16 | John  ...   | 2009-11-18 22:26 | Install | 2
   15 | John  ...   | 2009-11-18 22:24 | Install | 3
   14 | John  ...   | 2009-11-18 22:22 | Install | 10
   13 | John  ...   | 2009-11-18 22:18 | Install | 10
   12 | John  ...   | 2009-11-18 22:15 | Install | 14
   11 | John  ...   | 2009-11-18 22:13 | Install | 1
   10 | John  ...   | 2009-11-18 22:13 | Install | 2
   9 | System    | 2009-11-18 22:10 | Update | 16

To list packages of a transaction, use option info followed by transaction ID:
# yum history info 28
Loaded plugins: downloadonly, fastestmirror, keys, presto, refresh-packagekit
Transaction ID : 28
Begin time : Fri Nov 20 22:55:05 2009
Begin rpmdb : 1688:e57164ff4d8c1577daaa7e4711a6bcd20014fdfc
End time : 22:56:02 2009 (57 seconds)
End rpmdb : 1700:bc600702696126eaee80165df7933e0c02b69aa5
User : System
Return-Code : Success
Transaction performed with:
   Installed rpm-4.7.1-6.fc12.i686
   Installed yum-3.2.25-1.fc12.noarch
   Installed yum-metadata-parser-1.1.2-14.fc12.i686
   Installed yum-plugin-fastestmirror-1.1.24-2.fc12.noarch
Packages Altered:
   Dep-Install aspell-12:0.60.6-7.fc12.i686
   Install bilbo-1.0-1.fc12.i686
   Install blogtk-1.1-13.fc12.noarch
   Dep-Install clucene-core-0.9.21-4.fc12.i686
   Dep-Install kde-filesystem-4-30.fc12.noarch
   Dep-Install kde-settings-4.3-12.noarch
   Dep-Install kdelibs-6:4.3.2-4.fc12.i686
   Dep-Install kdelibs-common-6:4.3.2-4.fc12.i686
   Dep-Install kdepimlibs-4.3.2-1.fc12.i686
   Dep-Install oxygen-icon-theme-4.3.2-1.fc12.noarch
   Dep-Install soprano-2.3.1-1.fc12.i686
   Dep-Install strigi-libs-0.7.0-1.fc12.i686

To undo a transaction, use option 'undo' with the transaction ID:
# yum history undo 28
Loaded plugins: downloadonly, fastestmirror, keys, presto, refresh-packagekit
Loading mirror speeds from cached hostfile
* fedora: ftp.nluug.nl
* rpmfusion-free: rpmfusion.famillecollet.com
* rpmfusion-free-updates: rpmfusion.famillecollet.com
* rpmfusion-nonfree: rpmfusion.famillecollet.com
* rpmfusion-nonfree-updates: rpmfusion.famillecollet.com
* updates: ftp.nluug.nl
Undoing transaction 28, from Fri Nov 20 22:55:05 2009
   Dep-Install aspell-12:0.60.6-7.fc12.i686
   Install bilbo-1.0-1.fc12.i686
   Install blogtk-1.1-13.fc12.noarch
   Dep-Install clucene-core-0.9.21-4.fc12.i686
   Dep-Install kde-filesystem-4-30.fc12.noarch
   Dep-Install kde-settings-4.3-12.noarch
   Dep-Install kdelibs-6:4.3.2-4.fc12.i686
   Dep-Install kdelibs-common-6:4.3.2-4.fc12.i686
   Dep-Install kdepimlibs-4.3.2-1.fc12.i686
   Dep-Install oxygen-icon-theme-4.3.2-1.fc12.noarch
   Dep-Install soprano-2.3.1-1.fc12.i686
   Dep-Install strigi-libs-0.7.0-1.fc12.i686

[...]

Remove 12 Package(s)

Reinstall 0 Package(s)

Downgrade 0 Package(s)

Is this ok [y/N]: y

[...]

Undoing transaction 28 creates a transaction 29. If you want to reinstall what was installed in transaction 28, do the following:

# yum history redo 28

Loaded plugins: downloadonly, fastestmirror, keys, presto, refresh-packagekit

Loading mirror speeds from cached hostfile

* fedora: ftp.nluug.nl

* rpmfusion-free: rpmfusion.famillecollet.com

* rpmfusion-free-updates: rpmfusion.famillecollet.com

* rpmfusion-nonfree: rpmfusion.famillecollet.com

* rpmfusion-nonfree-updates: rpmfusion.famillecollet.com

* updates: ftp.nluug.nl

Repeating transaction 28, from Fri Nov 20 22:55:05 2009

Dep-Install aspell-12:0.60.6-7.fc12.i686

Install bilbo-1.0-1.fc12.i686

Install blogtk-1.1-13.fc12.noarch

Dep-Install clucene-core-0.9.21-4.fc12.i686

Dep-Install kde-filesystem-4-30.fc12.noarch

Dep-Install kde-settings-4.3-12.noarch

Dep-Install kdelibs-6:4.3.2-4.fc12.i686

Dep-Install kdelibs-common-6:4.3.2-4.fc12.i686

Dep-Install kdepimlibs-4.3.2-1.fc12.i686

Dep-Install oxygen-icon-theme-4.3.2-1.fc12.noarch

Dep-Install soprano-2.3.1-1.fc12.i686

Dep-Install strigi-libs-0.7.0-1.fc12.i686

[...]

Install 12 Package(s)

Upgrade 0 Package(s)

Total download size: 35 M

Is this ok [y/N]: y

[...]

Option 'summary' associated with a package name gives the list of transaction where that package is involved.

# yum history summary blogtk

Loaded plugins: downloadonly, fastestmirror, keys, presto, refresh-packagekit

---------------------------------------------------------------------

John Smith | Last day | E, I | 24

System | Last day | Install | 12

And option 'list'

# yum history list blogtk

Loaded plugins: downloadonly, fastestmirror, keys, presto, refresh-packagekit

ID | Login user | Date and time | Action(s) | Altered

---------------------------------------------------------------------------

30 | John ... | 2009-11-20 23:41 | Install | 12

29 | John ... | 2009-11-20 23:30 | Erase | 12

28 | System | 2009-11-20 22:55 | Install | 12

And option 'info' associated with package name gives a full information of transactions where it was involved.

# yum history info blogtk

Loaded plugins: downloadonly, fastestmirror, keys, presto, refresh-packagekit

Transaction ID : 29

Begin time : Fri Nov 20 23:30:34 2009

Begin rpmdb : 1700:bc600702696126eaee80165df7933e0c02b69aa5

End time : 23:30:56 2009 (22 seconds)

End rpmdb : 1688:e57164ff4d8c1577daaa7e4711a6bcd20014fdfc

User : John Smith

Return-Code : Success

Transaction performed with:

Installed rpm-4.7.1-6.fc12.i686

Installed yum-3.2.25-1.fc12.noarch

Installed yum-metadata-parser-1.1.2-14.fc12.i686

Installed yum-plugin-fastestmirror-1.1.24-2.fc12.noarch

Packages Altered:

Erase aspell-12:0.60.6-7.fc12.i686

Erase bilbo-1.0-1.fc12.i686

Erase blogtk-1.1-13.fc12.noarch

Erase clucene-core-0.9.21-4.fc12.i686

Erase kde-filesystem-4-30.fc12.noarch

Erase kde-settings-4.3-12.noarch

Erase kdelibs-6:4.3.2-4.fc12.i686

Erase kdelibs-common-6:4.3.2-4.fc12.i686

Erase kdepimlibs-4.3.2-1.fc12.i686

Erase oxygen-icon-theme-4.3.2-1.fc12.noarch

Erase soprano-2.3.1-1.fc12.i686

Erase strigi-libs-0.7.0-1.fc12.i686

-------------------------------------------------------------------------------

Transaction ID : 28

Begin time : Fri Nov 20 22:55:05 2009

Begin rpmdb : 1688:e57164ff4d8c1577daaa7e4711a6bcd20014fdfc

End time : 22:56:02 2009 (57 seconds)

End rpmdb : 1700:bc600702696126eaee80165df7933e0c02b69aa5

User : System

Return-Code : Success

Transaction performed with:

Installed rpm-4.7.1-6.fc12.i686

Installed yum-3.2.25-1.fc12.noarch

Installed yum-metadata-parser-1.1.2-14.fc12.i686

Installed yum-plugin-fastestmirror-1.1.24-2.fc12.noarch

Packages Altered:

Dep-Install aspell-12:0.60.6-7.fc12.i686

Install bilbo-1.0-1.fc12.i686

Install blogtk-1.1-13.fc12.noarch

Dep-Install clucene-core-0.9.21-4.fc12.i686

Dep-Install kde-filesystem-4-30.fc12.noarch

Dep-Install kde-settings-4.3-12.noarch

Dep-Install kdelibs-6:4.3.2-4.fc12.i686

Dep-Install kdelibs-common-6:4.3.2-4.fc12.i686

Dep-Install kdepimlibs-4.3.2-1.fc12.i686

Dep-Install oxygen-icon-theme-4.3.2-1.fc12.noarch

Dep-Install soprano-2.3.1-1.fc12.i686

Dep-Install strigi-libs-0.7.0-1.fc12.i686

Wrong predictions are so funny.

2009-06-13T21:38:00.001+01:00

According to this website listing the top 30 failed technology predictions, these people said:

“There is no reason anyone would want a computer in their home.” — Ken Olson, president, chairman and founder of Digital Equipment Corp. (DEC), maker of big business mainframe computers, arguing against the PC in 1977.
“We will never make a 32 bit operating system.” — Bill Gates

Well, I have a 32-bit laptop and I am at home writing this note.

SSH tunneling 4

2009-05-29T08:37:00.002+01:00

mdomsch left a comment on a previous post. So here is a 4th way to create a tunnel.
In the file $HOME/.ssh/config, just add the following lines:
Host serverC
ProxyCommand ssh serverB nc serverC 22

nc is the netcat utility that you have to install with yum on serverB:
$ yum install nc

My serverB is running under SUSE Linux Enterprise Server 10. The package I had to install and the binary are called netcat. So I have a line like this in $HOME/.ssh/config:
ProxyCommand ssh serverB netcat serverC 22

Then just type this and answer 'yes':

[serverA]$ ssh serverC
No protocol specified
Warning: No xauth data; using fake authentication data for X11 forwarding.
The authenticity of host 'serverC (<no hostip for proxy command>)' can't be established.
RSA key fingerprint is d8:d7:b6:f7:a2:4d:7c:e5:b4:d3:18:99:ca:b2:0f:e4.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added 'serverC' (RSA) to the list of known hosts.
Warning: No xauth data; using fake authentication data for X11 forwarding.

When leaving your session on serverC, you will notice
Killed by signal 1.
It appears also when using scp and sftp. It's no big deal actually.

This last post is my favorite solution for creating an SSH tunneling. Once your .ssh/config file is written, you only need to remember the name of the server you want to connect (here serverC).

Usefull SSH tools

2009-05-28T16:37:00.002+01:00

I have many desktops and laptops to manage in my office. They are all running under fedora 10 and I was getting tired to repeat the same commands on each client.
So I found two nice little tools that launches the same small jobs on a all my clients.

These tools are pssh (Parallel SSH) and mussh (Multiple SSH).

You can install the packages with yum:
# yum install pssh mussh

pssh is a set of 5 python scripts:

pssh is used to execute a command on several hosts,
pscp is used to copy a local file to remote hosts,
pslurp is used to copy a file from remote hosts onto the local host,
prsync runs an rsync command, does essentially the same as pscp,
pnuke kills all processes that match a pattern.

The pssh script (and the 4 others scripts) runs in parallel while mussh runs sequentially.

With mussh:
$ time mussh -H hosts.cfg -c 'rpm -qa --qf "%{SIZE}:%{NAME}:%{ARCH}\n" | sort -n > /dev/null'
real 1m39.577s

and with pssh:
$ time pssh -h hosts.cfg 'rpm -qa --qf "%{SIZE}:%{NAME}:%{ARCH}\n" | sort -n > /dev/null'
[1] 15:55:24 [SUCCESS] host1 22
[2] 15:55:24 [SUCCESS] host2 22
[3] 15:55:24 [SUCCESS] host3 22
[4] 15:55:24 [SUCCESS] host4 22
[5] 15:55:24 [SUCCESS] host5 22
[6] 15:55:24 [SUCCESS] host6 22
[7] 15:55:28 [SUCCESS] host7 22
[8] 15:55:30 [SUCCESS] host8 22
[9] 15:55:30 [SUCCESS] host9 22
[10] 15:55:30 [SUCCESS] host10 22
[11] 15:55:30 [SUCCESS] host11 22
[12] 15:55:31 [SUCCESS] host12 22
real 0m12.838s

which gives a speedup of 8.

hosts.cfg is a file containing a list of hosts, one host per line.
$ cat hosts.cfg
host1
host2
host3
host4
host5
host6
host7
host8
host9
host10
host11
host12

What I think is missing in pssh is the possibility to enter with the option -h on the command line the list of hosts on which you want to run your command.
$ mussh -h host1 host2 host5 -c 'rpm -qa --qf "%{SIZE}:%{NAME}:%{ARCH}\n" | sort -n > /dev/null'
$ mussh -H hosts.cfg -c 'rpm -qa --qf "%{SIZE}:%{NAME}:%{ARCH}\n" | sort -n > /dev/null'

With pscp, you can add 'localhost:22200' in hosts.cfg if you created an SSH tunnel (See my previous posts about ssh tunneling).

SSH tunneling 3

2009-05-28T14:03:00.001+01:00

The following is not really ssh tunneling but it does the same thing.

server A ------> server B ------> server C
| |
+---------------------------------+

If you want to connect to server C by typing one single command line, you would write:
[serverA]$ ssh serverB -t ssh serverC
[serverC]$

You won't be able to use scp to copy a file from server A to server C. If you want, check here and here and then do:
[server A]$ scp -P 22200 myfile.txt localhost:

SSH tunneling 2

2009-05-27T14:33:00.001+01:00

Following my previous post about ssh tunneling, I found a little GUI tool called gSTM (Gnome Ssh Tunnel Managment).

Once the tunnel is running, you just need to to run in a terminal the following command:
[serverA]$ ssh -p 22200 localhost
and you are connected to server C.
gSTM is run in the background. You will see the gSTM icon in the notification area:

SSH tunneling

2009-05-27T09:06:00.009+01:00

At office, I need to do 2 ssh connections to reach the server I want to work on. And usually, I need several terminals connected to that server. So I first connect to server B from server A, and then I connect to server C from server B.
My dream was to type one single ssh command on server A and to get myself connected to server C.

server A ------> server B ------> server C
| |
+---------------------------------+

On server A, type the following in a terminal:
[serverA]$ ssh -L22200:localhost:22200 serverB

You are now connected on server B. Type the following in the same terminal:
[serverB]$ ssh -L22200:localhost:22 serverC

You are now connected to server C.

You have create a tunnel that goes from server A to server C.
Now, open a new terminal and type the following:
[serverA]$ ssh -p 22200 localhost

You are now connected to server C.
I used here port 22200. Any other unused port can be used.

Are you more Leonidas or Leonidas?

2009-03-11T18:03:00.001+01:00

Everytime people are talking about , I think of, the famous belgian chocolate.

Video projector DELL M109S + Fedora: Did somebody succeed to make it work?

2009-03-02T10:38:00.002+01:00

I just received the DELL video projector M109S. It is one of the smallest one I've ever seen. And it is using LEDs as a light source. Sounds very ecological. Even more since I cannot make it work with my DELL Inspiron 9400 laptop running Fedora 10. Pressing the keys Fn+F8 (CRT/LCD) doesn't work. Works well with Windows though.
Did anybody succeed to make it work properly under Fedora?

F10: How to fix the getpwuid_r() failure?

2009-02-12T14:06:00.001+01:00

You may have noticed that running acroread gives you a warning. It is actually an error because acroread doesn't start at all.
(acroread:18426): GLib-WARNING **: getpwuid_r(): failed due to unknown user id (1010)
This happens because you are using LDAP to authenticate yourself on your machine.

The fix is veru easy. Just start the nscd service.

/etc/init.d/nscd start
or
service nscd start

and make sure it will start during the next reboot:

chkconfig nscd on

F10: How to setup a pxeboot + tftp + dhcp server?

2008-12-05T12:58:00.009+01:00

I want to be able to boot over the network to install Fedora 10 on several machines at the same time.
And I have a hardware firewall, a Netscreen NS50, which has a dhcp server configured in it.

The solution is to confiure a tftp server (so the client will be able to boot on through PXE), a dhcp server different from the one in my firewall and an NFS server.
The NFS server is called nfsserver (10.0.21.17), the dhcp and the tftp server are on the same server, called pxeserver (10.0.21.210). These server can reside on different machines.

Here are the packages to be installed:

system-config-netboot
system-config-netboot-cmd
tftp
tftp-server
dhcpd

Configuration of the NFS server on nfsserver
Download the DVD iso image of the Fedora 10 distribution.
Then mount it in a directory /mnt/f10tmp:
# mkdir /mnt/f10tmp
# mount Fedora-10-i386-DVD.iso /mnt/f10tmp -o loop

Copy the content of /mnt/f10tmp into /mnt/f10.
# mkdir /mnt/f10
# cp -r /mnt/f10tmp/* /mnt/f10

Add the following line in /etc/exports:
/mnt/f10 *(rw,async,no_root_squash)

Then run
# exportfs -va

Configuration of the tftp server on pxeserver
Run system-config-netboot and configure a network installation:

Oprating system identifier: name of your distribution (Fedora10)
Description: some desciption!
Select protocol for installation: choose NFS
Kickstart: if you have a kickstart file, put the path to it here, if not, leave empty
Server IP address : 10.0.21.17
Location: /mnt/f10

Click OK. This creates a directory /tftpboot/linux-install/Fedora10 with files like initrd.img and vmlinuz downloaded from your nfsserver (/mnt/f10/images/pxeboot/).

Create a New PXE configuration file:

Hostname of IP Address/subnet: the IP of your hosts
Operating system: Fedora10 should be there. It is auto-populated from the previous step.

Click OK. A file named default is created in /tftpboot/linux-install/pxelinux.cfg.

Edit the file /etc/xinetd.d/tftp and change the line containing server_args:

service tftp
{
disable = no
socket_type = dgram
protocol = udp
wait = yes
user = root
server = /usr/sbin/in.tftpd
server_args = -s /tftpboot -v -v
per_source = 11
cps = 100 2
flags = IPv4
}

Check that the server is on:
# chkconfig --list tftp
tftp on

If off, set it on:
# chkconfig tftp on

and the start the server xinetd:
# /etc/init.d/xinetd restart

You can test your tftp server with the tftp client from your home directory:
$ tftp pxeserver
get linux-install/pxelinux.0

You should have the file pxelinux.0 in your home directory.

Configuration of the dhcp server on pxeserver
Edit the file /etc/dhcpd.conf and add the following lines:

subnet 10.0.21.208 netmask 255.255.255.240 {
range 10.0.21.209 10.0.21.222;
allow booting;
allow bootp;
class "pxeclients" {
match if substring(option vendor-class-identifier, 0, 9) = "PXEClient";
next-server 10.0.21.210;
filename "linux-install/pxelinux.0";
}
}

Even if you have a running dhcp server somewhere, this will work.

Start the dhcp server:
# /etc/init.d/dhcpd restart

Check in /var/log/messages:
Dec 5 11:49:06 pxeserevr dhcpd: Listening on LPF/eth0/00:15:c5:53:fd:f5/10.0.21.208/28
Dec 5 11:49:06 pxeserver dhcpd: Sending on LPF/eth0/00:15:c5:53:fd:f5/10.0.21.208/28
Dec 5 11:49:06 pxeserver dhcpd: Sending on Socket/fallback/fallback-net

Restart now your client (target) pc to boot over the network.

You should see something like this in /var/log/messages:

Dec 5 11:49:32 pxeserver dhcpd: DHCPDISCOVER from 00:04:76:51:2e:d8 via eth0
Dec 5 11:49:33 pxeserver dhcpd: DHCPOFFER on 10.0.21.212 to 00:04:76:51:2e:d8 via eth0
Dec 5 11:49:36 pxeserver dhcpd: DHCPREQUEST for 10.0.21.212 (10.0.21.210) from 00:04:76:51:2e:d8 via eth0
Dec 5 11:49:36 pxeserver dhcpd: DHCPACK on 10.0.21.212 to 00:04:76:51:2e:d8 via eth0
Dec 5 11:49:36 pxeserver in.tftpd[6567]: RRQ from 10.0.21.212 filename linux-install/pxelinux.0
Dec 5 11:49:36 pxeserver in.tftpd[6567]: tftp: client does not accept options

If it does not work, either disable the firewall (system-config-firewall) on pxeserver or allow port 69 to be accessible to any hosts accessing the tftp server.
If you get an error, please have a look here to fix it.

You will have a prompt to boot either on Fedora10 or on your local installation.

F10: Preupgrade from F9 to F10

2008-11-27T16:18:00.004+01:00

Last may, I upgraded to Fedora 9 using the preupgrade tool. And it went well. So for Fedora 10, it could only go better!
I patient until a new version of preupgrade for F9 (1.0.0-1) was released this morning. I then ran preupgrade, chose Fedora 10 (Cambridge) and indeed, everything went almost well.

Preupgrade found out I was running out of disk space after downloading all the f10 packages, so I cancelled the preupgrade, remove some packages like openoffice-org to make sure I had enough space on disk. Then I re-ran preupgrade, all the packages have been installed except those for NVidia.

After rebooting, the X11 server ran with the nouveau driver instead of the nvidia one. But that's fine. I'll install the nvidia driver later.

And remember to have enough free space on your /boot partition (500 Mb is comfortable).

F9: How to authenticate using a usb flash drive?

2008-10-14T16:21:00.004+01:00

I recently stumbled upon the PAM_USB project.
Using a small plugin called pam_usb.so, you are able to log in to your linux pc without typing your password by just using a usb key. I found it is a very nice way to recycle your old usb key of several MB since the files stored for the authentication have sizes of about 1kb.

First install the rpm package (see below)
insert your usb key
then run the following: pamusb-conf --add-device MyUsbDevice
This will create some configuration in the file /etc/pamusb.conf.
then run this to create configurations for user john:

# pamusb-conf --add-user john
Which device would you like to use for authentication ?
* Using "MyUsbDevice" (only option)

User : john
Device : MyUsbDevice

Save to /etc/pamusb.conf ?
[Y/n]
Done.

You can do it also for root.
Once this is done, you need to add the following line in the file /etc/pam.d/system-auth:
auth sufficient pam_usb.so
above the line:
auth sufficient pam_unix.so nullok try_first_pass
Other files in /etc/pam.d can get this line, for example, gdm or yumex (for root).
In the gdm file, if your screensaver locks your system, you won't need to type your password. In the yumex file, you don't need to type the root password usually asked. You'll be able to connect as root with the "su -" with typing any password.

$ su -
* pam_usb v0.4.2
* Authentication request for user "root" (su-l)
* Device "MyUsbDevice" is connected (good).
* Performing one time pad verification...
* Access granted.
#
If you remove your usb key, you will still be asked for your password.

IMPORTANT: If you leave your pc for some time and lock your screen manually, don't forget to remove your usb key!

Also if someone connects to your pc with your usb key plugged, he won't be able to connect as root with "su -"!

I found a spec file that I updated. Just recompile the rpm package with "rpmbuild -ba pam_usb.spec". You need to install package rpmdevtools.

Name: pam_usb
Summary: Authentication using ordinary USB Flash Drives
Version: 0.4.2
Release: 1%{?dist}
License: GPL
Group: System Environment/Base
URL: http://www.pamusb.org
Source0: http://download.sourceforge.net/pamusb/%{name}-%{version}.tar.gz
Source4: pam_usb-pam.d-login.example
Source5: pam_usb-pam.d-gdm.example

BuildRoot: %{_tmppath}/%{name}-%{version}-%{release}-root-%(%{__id_u} -n)
BuildRequires: pam-devel, hal-devel, pygobject2-devel, libxml2-devel
Requires: pygobject2
%if "%{?fedora}" < "7"
BuildRequires: python-elementtree
Requires: python-elementtree
%endif

%description
pam_usb provides hardware authentication for Linux using ordinary
USB Flash Drives. It works with any application supporting PAM,
such as su, any login manager (GDM, KDM), etc. The pam_usb package contains:

* A PAM Module
o Password-less authentication. Use your USB stick for authentication,
don’t type passwords anymore.
o Device auto probing.You don’t need to mount the device, or even
to configure the device location (sda1, sdb1, etc).
pam_usb.so will automatically locate the device using HAL and access
its data by itself.
o Two-factor authentication. Achieve greater security by requiring
both the USB stick and the password to authenticate the user.
o Non-intrusive. pam_usb doesn’t require any modifications of the
USB storage device to work (no additional partitions required).
o USB Serial number, model and vendor verification.
o Support for One Time Pads authentication.
* Several tools
o pamusb-agent: trigger actions (such as locking the screen)
upon device authentication and removal.
o pamusb-conf: configuration helper.
o pamusb-check: integrate pam_usb’s authentication engine within
your scripts or applications.

%prep
%setup -q

mkdir pam.d-examples
install -p -m644 %{SOURCE4} pam.d-examples/login
install -p -m644 %{SOURCE5} pam.d-examples/gdm

# fix lib64
sed -i -e 's|/lib/security|/%{_lib}/security|' Makefile

%build
# Note: we may not overwrite CFLAGS, so use CC instead...
make CC="%__cc $RPM_OPT_FLAGS" LIB=%{_lib} %{?_smp_mflags}

%install
rm -rf $RPM_BUILD_ROOT

make install LIB=%{_lib} DESTDIR=$RPM_BUILD_ROOT

# marked these as %doc already
rm -rf $RPM_BUILD_ROOT/%{_docdir}/pamusb/

%clean
rm -rf $RPM_BUILD_ROOT

%files
%defattr(-,root,root,-)
/%{_lib}/security/*
%{_bindir}/pamusb*
%config(noreplace) %{_sysconfdir}/pamusb.conf
%doc COPYING ChangeLog pam.d-examples doc/QUICKSTART doc/CONFIGURATION doc/FAQ
%attr(644,root,root) %{_mandir}/*/pamusb*

%changelog
* Sat Aug 25 2007 John Mahowald - 0.4.1-1
- 0.4.1

* Sat May 5 2007 John Mahowald - 0.4.0-1
- 0.4.0
- drop hotplug subpackage

* Fri Sep 1 2006 Dmitry Butskoy - 0.3.3-6
- rebuild for FC6

* Tue Aug 1 2006 Dmitry Butskoy - 0.3.3-5
- avoid world-writable docs (#200840)

* Wed Feb 15 2006 Dmitry Butskoy - 0.3.3-4
- rebuild for FC5

* Thu Feb 2 2006 Dmitry Butskoy - 0.3.3-3
- pam_usb-hotplug requires udev now instead of the obsoleted hotplug

* Thu Dec 29 2005 Dmitry Butskoy - 0.3.3-2
- accepted for Fedora Extras (review by Wart )

* Thu Dec 22 2005 Dmitry Butskoy - 0.3.3-2
- Use /etc/udev/rules.d instead of /etc/hotplug.d
- Add some pam.d examples (Wart@kobold.org)

* Tue Dec 20 2005 Dmitry Butskoy - 0.3.3-1
- upgrade to 0.3.3
- always build hotplug package
- add Makefile patch (directory creation, permissions etc.)
- create usbhotplug pam file own way

* Wed Sep 7 2005 Dmitry Butskoy - 0.3.2-1
- initial release
- copy quckstart.html and options.html from web-site to package
- separate hotplug stuff as an additional package, build it
when "--with hotplug" is specified only

F9: How to install Fedora from a usb stick (or any other hard disk)

2008-09-30T13:45:00.000+01:00

The following steps suppose that your pc can boot on a usb device:

Insert the usb stick and note which device is used:
# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/mapper/VolGroup01-LogVol00
                      11062528   9284412   1778116 84% /
/dev/mapper/VolGroup01-LogVol01
                     101000512 79661560 16125564 84% /home
/dev/sda2               505636     30011    449520   7% /boot
tmpfs                  1036992       420   1036572   1% /dev/shm
/dev/sdb1              4032804     11896   4020908   1% /media/disk

In this case /dev/sdb1 contains Fedora DVD iso. Note this down as you need it later.
Copy the Fedora DVD iso to directory /media/disk.
# cp Fedora-9-i386-DVD.iso /media/disk
Unmont the /media/disk directory:
# umount /media/disk
Use the iso image of Fedora 9 i386 (DVD) to make a bootable usb stick (more details here):
livecd-iso-to-disk --noverify Fedora-9-i386-DVD.iso /dev/sdb1
Remove the usb stick and re-insert it to check its content.
Step 3 has created 2 directories, LiveOS and syslinux

# ls /media/disk/*
/media/disk/LiveOS:

/media/disk/syslinux:
boot.cat boot.msg general.msg grub.conf initrd.img isolinux.bin ldlinux.sys memtest options.msg param.msg rescue.msg splash.jpg syslinux.cfg trans.tbl vesamenu.c32 vmlinuz
Reboot
Reboot your system and boot on the usb device.
Install from hard disk
While in installation wizard, select "Hard drive" as installation method and choose /dev/sdb1 as it contains the Fedora DVD iso. And rest is damn easy.

If your pc cannot boot on a usb device, there is another way. Replace steps 2, 3, 4 with the following:

Mount Fedora DVD iso on /mnt/
# mount Fedora-9-i386-DVD.iso /mnt/ -ro loop
Copy the initrd.img and vmlinuz to /boot/ partition
# cd /mnt/isolinux/
# cp initrd.img vmlinuz /boot/
Create grub entry for booting into Fedora 9
Add these lines at the end of your /etc/grub.conf file.

title Fedora 9 (New installation)
kernel /vmlinuz
initrd /initrd.img

Reboot your system and boot into the Fedora 9 (New installation) grub entry.

Red Hat is not what you think it is!

2008-08-20T10:57:00.005+01:00

This picture was taken in a restaurant in Japan.

Ethernet bonding

2008-07-01T09:03:00.003+01:00

Ethernet bonding is a way to combine multiple ethernet channel into a single one. If, for example, you have 2 ethernet cards of 1 Gb/s speed each, you can combine them to have one virtual card of 2 Gb/s speed.

Suppose you have a cluster of several nodes, one of them is the master node and has 3 ethernet cards, one for the network 192.168.1.0/24 (file ifcfg-eth2) and the two others (file ifcfg-eth0 and ifcfg-eth1) bonded together (file ifcfg-bond0) for the network 192.168.7.0/24.

The ethernet card in the network should be configured with file ifcfg-eth2 as follows:
DEVICE=eth2
BOOTPROTO=none
HWADDR=00:05:48:A8:F2:H6
ONBOOT=yes
TYPE=Ethernet
USERCTL=no
IPV6INIT=no
PEERDNS=yes
NETMASK=255.255.255.0
IPADDR=192.168.1.13
GATEWAY=192.168.1.254

Juste create in /etc/sysconfig/network-scripts, the file ifcfg-bond0 containing the following lines:
DEVICE=bond0
BOOTPROTO=none
BROADCAST=192.168.7.255
IPADDR=192.168.7.20
NETMASK=255.255.255.0
NETWORK=192.168.7.0
GATEWAY=192.168.1.254
ONBOOT=yes
USERCTL=no
TYPE=Ethernet

which is what you would have for a single ethernet card in a file called ifcfg-eth0. On the master node, you have to specify the gateway 192.168.1.254.

On the other nodes, you will specify the gateway to be 192.168.7.20 which is the IP address of the master node in the network 192.168.7.0/24:

DEVICE=bond0
BOOTPROTO=none
BROADCAST=192.168.7.255
IPADDR=192.168.7.22
NETMASK=255.255.255.0
NETWORK=192.168.7.0
ONBOOT=yes
USERCTL=no
TYPE=Ethernet
GATEWAY=192.168.7.20

You must also create the files ifcfg-eth0 and ifcfg-eth1, one file for each card in which you will put the following:

In ifcfg-eth0:
DEVICE=eth0
ONBOOT=yes
USERCTL=no
MASTER=bond0
SLAVE=yes
BOOTPROTO=no
TYPE=Ethernet

In ifcfg-eth1:
DEVICE=eth1
ONBOOT=yes
USERCTL=no
MASTER=bond0
SLAVE=yes
BOOTPROTO=no
TYPE=Ethernet

This way, both cards know who is the master.

On the master node, ifconfig should give:
# ifconfig
bond0 Link encap:Ethernet HWaddr 00:02:55:C7:ED:73
inet addr:192.168.7.20 Bcast:192.168.7.255 Mask:255.255.255.0
inet6 addr: fe80::200:ff:fe00:0/64 Scope:Link
UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
RX packets:27451 errors:0 dropped:0 overruns:0 frame:0
TX packets:13703 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:2542289 (2.4 MiB) TX bytes:1328584 (1.2 MiB)

eth0 Link encap:Ethernet HWaddr 00:02:55:C7:ED:73
inet6 addr: fe80::202:55ff:fec7:ed73/64 Scope:Link
UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
RX packets:13684 errors:0 dropped:0 overruns:0 frame:0
TX packets:6854 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:1239023 (1.1 MiB) TX bytes:663919 (648.3 KiB)

eth1 Link encap:Ethernet HWaddr 00:02:55:C7:ED:73
inet6 addr: fe80::202:55ff:fec7:ed73/64 Scope:Link
UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
RX packets:13767 errors:0 dropped:0 overruns:0 frame:0
TX packets:6849 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:1303266 (1.2 MiB) TX bytes:664665 (649.0 KiB)

eth2 Link encap:Ethernet HWaddr 00:03:47:B1:F9:E6
inet addr:192.168.1.13 Bcast:192.168.1.15 Mask:255.255.255.240
inet6 addr: fe80::203:47ff:feb1:f9e6/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:311014 errors:0 dropped:0 overruns:0 frame:0
TX packets:443837 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:110699104 (105.5 MiB) TX bytes:434496718 (414.3 MiB)
Memory:effc0000-effe0000

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:185354 errors:0 dropped:0 overruns:0 frame:0
TX packets:185354 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:421703302 (402.1 MiB) TX bytes:421703302 (402.1 MiB)

On the slave nodes, ifconfig should give
# ifconfig
bond0 Link encap:Ethernet HWaddr 00:02:55:C7:ED:EE
inet addr:192.168.7.22 Bcast:192.168.7.255 Mask:255.255.255.0
inet6 addr: fe80::202:55ff:fec7:edee/64 Scope:Link
UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
RX packets:14694 errors:0 dropped:0 overruns:0 frame:0
TX packets:3973 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:1138254 (1.0 MiB) TX bytes:359734 (351.3 KiB)

eth0 Link encap:Ethernet HWaddr 00:02:55:C7:ED:EE
inet6 addr: fe80::202:55ff:fec7:edee/64 Scope:Link
UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
RX packets:7380 errors:0 dropped:0 overruns:0 frame:0
TX packets:1991 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:579813 (566.2 KiB) TX bytes:178136 (173.9 KiB)
Interrupt:193 Base address:0x6000

eth1 Link encap:Ethernet HWaddr 00:02:55:C7:ED:EE
inet6 addr: fe80::202:55ff:fec7:edee/64 Scope:Link
UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
RX packets:7322 errors:0 dropped:0 overruns:0 frame:0
TX packets:1994 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:558969 (545.8 KiB) TX bytes:183558 (179.2 KiB)
Interrupt:201 Base address:0xa000

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:152 errors:0 dropped:0 overruns:0 frame:0
TX packets:152 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:10980 (10.7 KiB) TX bytes:10980 (10.7 KiB)

F9: How to configure compiz-fusion

2008-06-02T09:51:00.004+01:00

First install the compiz-fusion-release package:

rpm -Uvh http://www.dfm.uninsubria.it/compiz/compiz-fusion-release-1-1.noarch.rpm

and install compiz-fusion:

compiz-0.7.5-10.fc9.i386

compizconfig-python-devel-0.7.5-6.fc9.i386

compiz-fusion-release-1-1.noarch

compiz-gnome-0.7.5-10.fc9.i386

compiz-manager-0.6.0-7.fc9.noarch

libcompizconfig-0.7.5-6.fc9.i386

compiz-fusion-plugins-extra-0.7.5-6.fc9.i386

compiz-bcop-0.7.5-6.fc9.noarch

libcompizconfig-devel-0.7.5-6.fc9.i386

compiz-devel-0.7.5-10.fc9.i386

compizconfig-backend-kconfig-0.7.5-6.fc9.i386

compiz-fusion-plugins-main-0.7.5-7.fc9.i386

compiz-fusion-plugins-unsupported-0.7.5-6.fc9.i386

compizconfig-backend-gconf-0.7.5-6.fc9.i386

compiz-kde-0.7.5-10.fc9.i386

compizconfig-python-0.7.5-6.fc9.i386

Then run fusion-icon and compiz --replace.

Here are a screenshot I made:

Compiz-fusion's rotating cube. The left face shows the recently released Big Buck Bunny movie.

F9: Nvidia driver available for X server 1.5

2008-05-29T14:03:00.002+01:00

Nvidia just released a new driver (version 173.14.05) with preliminary support for X.Org server 1.5 (http://www.nvnews.net/vbulletin/showthread.php?t=113919). Livna has already done the packaging. Just install the following:

xorg-x11-drv-nvidia-libs.i386
kmod-nvidia.i686
kmod-nvidia-2.6.25.3-18.fc9.i686.i686
xorg-x11-drv-nvidia.i386

And it works! :-)

F9: Trouble using the nvidia driver

2008-05-22T07:48:00.005+01:00

For the first time since I am using Red Hat 7.3 (2002) and now Fedora, I could not use the Nvidia driver.

My graphic card is:

01:00.0 VGA compatible controller: nVidia Corporation GeForce Go 7900 GS (rev a1)

I struggled for hours trying to make it work, reading here and there, trying to configure my config file /etc/X11/xorg.conf correctly. Nothing works.

When I run the following:

startx -- :2 -config xorg.conf.nvidia -ignoreABI

I get these error messages, complaining about libwfb.so:

(II) Loading /usr/lib/xorg/modules/extensions/nvidia//libwfb.so
dlopen: /usr/lib/xorg/modules/extensions/nvidia//libwfb.so: undefined symbol: miZeroLineScreenIndex
(EE) Failed to load /usr/lib/xorg/modules/extensions/nvidia//libwfb.so
(II) UnloadModule: "wfb"
(EE) Failed to load module "wfb" (loader failed, 7)
(II) Loading sub module "ramdac"
(II) LoadModule: "ramdac"
(II) Module "ramdac" already built-in
(II) resource ranges after probing:
[0] -1 0 0xffffffff - 0xffffffff (0x1) MX[B]
[1] -1 0 0x000f0000 - 0x000fffff (0x10000) MX[B]
[2] -1 0 0x000c0000 - 0x000effff (0x30000) MX[B]
[3] -1 0 0x00000000 - 0x0009ffff (0xa0000) MX[B]
[4] -1 0 0x0000ffff - 0x0000ffff (0x1) IX[B]
[5] -1 0 0x00000000 - 0x00000000 (0x1) IX[B]
(**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32
(==) NVIDIA(0): RGB weight 888
(==) NVIDIA(0): Default visual is TrueColor
(==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
(**) NVIDIA(0): Option "AddARGBGLXVisuals" "True"
(**) NVIDIA(0): Option "OnDemandVBlankInterrupts" "True"
(**) NVIDIA(0): Enabling RENDER acceleration
(II) NVIDIA(0): Support for GLX with the Damage and Composite X extensions is
(II) NVIDIA(0): enabled.
(EE) NVIDIA(0): Failed to load the NVIDIA kernel module!
(EE) NVIDIA(0): *** Aborting ***
(II) UnloadModule: "nvidia"
(II) UnloadModule: "fb"
(EE) Screen(s) found, but none have a usable configuration.

Fatal server error:
no screens found

I even checked in the Release Notes and found nothing about Nvidia driver not working in F9.

Fedora should have written in the Release Notes that the X server is not compatible with the latest nvidia driver 173.08-1 (as said here)

Anyway, let's wait for the next release of Xorg and nvidia driver.

---------
PS: I found a link explaing how to downgrade xorg and make everything work. I did not try.

F9: Preupgrade from F8 to F9 (Success!)

2008-05-15T07:50:00.006+01:00

So I used to preupgrade (0.9.3-3) to upgrade from F8 to F9. Version 0.9.3-1 was a bit buggy.
All the F9 package will be installed in /var/cache/yum/anaconda-upgrade. The boot images will be downloaded in /boot so if you created a special partition for /boot, be aware that if you don't have enough space in it, preupgrade will fail.
I suggest that you remove all the unnecessary kernels that you have installed previoulsy. You probably need something like 114 Mb (118413565 bytes to be precise). Fortunately, I had 500 Mb. :-)
These images are downloaded in /boot/upgrade.
After everything has been downloaded, you will need to reboot and choose in the grub menu Upgrade to Fedora 9 (Sulphur).
My laptop was connected on the network and almost at the end I had to configure my network card. I don't know if the preupgrade will work without a network connection.
After everything was installed, I rebooted again and selected Fedora (2.6.25-14.fc9.i686) in the grub menu. Everything went fine except that NetworkManager (NM) does not connect the PC to the network. I have to manually un-plug and then re-plug the cable so NM detects it and the network card gets an IP from the dhcp server.

F9: How to live upgrade from F8 to F9?

2008-05-13T11:05:00.002+01:00

Well, I used to use yum to perform a live update (see here) but from what I read today (Interview of Seth Vidal and Will Woods), I think I will try the PreUpgrade tool.
It can be downloaded from Koji's repository.

F9 accepted feature list

2008-04-16T17:33:00.013+01:00

Here is the list of the features that we will have in the next release of F9 (sulphur). I'm looking forward to install and play with it.

Bluetooth Ongoing Bluetooth enhancements

Clock Applet Improve the support for multiple timezones/locations in the clock applet

Dictionary Support Fix the proliferation of dictionaries in Fedora

Full Virt Kernel Boot Xen fully virtualized guests will be able to directly boot a kernel and initrd image and pass kernel boot args

ext4 filesystem Add ext4 filesystem support

Encrypted Filesystems Add built in encrypted file system support

Fast X start up and shutdown Increase response time to 1 second from exec to accepting clients

Firefox 3 Upgrade to Firefox 3

Free IPA Make Identity, Policy and Audit centrally and more easily managed

gcc4.3 Update Fedora to gcc 4.3

gdm Update to new upstream Gnome Desktop Manager (gdm)

Gvfs Gvfs is a userspace virtual filesystem with backends for things like sftp, ftp, dav, smb, obexftp, etc. It is the replacement/successor of gnome-vfs

jigdo Provide jigdo support for obtaining Fedora

KDE4 Replace KDE3.5 with 4.0 and upgrade modules

Live Persistence Live USB images persist between booting

More NetworkManager Expand the situations in which NetworkManager can be used and make it more robust

NetworkManager Mobile Broadband Proper support of GSM and CDMA mobile broadband cards in NetworkManager, including first-class device support like Ethernet and WiFi devices

PackageKit Provide alternative package management system

Partition Resizing Partition resizing at install time

perl5.10 Update perl in Fedora 9 to 5.10.0

Pre Upgrade Allow users to depsolve and download only the pkgs they need to upgrade their distro while using the older version

Randr Support Use the capabilities offered by the Xrandr X extension throughout the desktop

Second Stage Install Source Move install selection to second stage of installer

TeXLive Replace teTeX with TeXLive

Update X Server Update X server to at least version 1.5

Upstart Replacement for SysVinit that allows service management and event-driven initialization

Virtual Authentication Extend remote authentication to use Kerberos and PAM

Virtual Manager PolicyKit Enable virt-manager to run as a unpriveledged user

Xen paravirt_ops Replacing the current forward-ported XenSource code in kernel-xen with the paravirt_ops based implementation from upstream

XULRunner Introducing and ensuring that all the applications that need only the Gecko browser engine are using XULRunner

Flying Linux ...

2008-04-08T08:09:00.002+01:00

You really want to have a look a this small beautiful video. You will feel happy for the rest of the day.

F8: How to stop a service when booting on a wifi network

2008-04-06T08:58:00.003+01:00

To avoid being cracked by some hacker when you are connected on a wifi network, you can stop the sshd service while logging with the followng scipt:

[root@localhost ~]# more /etc/NetworkManager/dispatcher.d/sshd-script
#!/bin/bash

if [ -x /usr/bin/logger ]; then
LOGGER="/usr/bin/logger -s -p user.notice -t NetworkManagerDispatcher"
else
LOGGER=echo
fi

if [ ! -x /etc/init.d/sshd ]; then
$LOGGER "init script /etc/init.d/sshd missing or not executible"
return
fi

if [ -n $1 ] && [ $1 == "wlan0" ] && [ $2 == "up" ]; then
if [ -f /var/run/sshd.pid ]; then
$LOGGER "Network device $1 ($2): sshd is running, stopping it."
/sbin/service sshd stop
fi
fi

F8: How to install Live images on a USB drive

2008-03-28T10:31:00.001+01:00

You decided to test Fedora 9-beta but you don't want to install it yet on your disk. A good solution is to install the live version of Fedora 9-beta on your usb stick and then boot on it.
But how to install an iso image on a usb drive?

Well, insert your usb drive and check on which device your usb drive is mounted.

# mount
[...]
/dev/sdb1 on /media/disk type vfat (rw,nosuid,nodev,uhelper=hal,shortname=lower,uid=501)

Then check that you have the necessary tools installed on your pc.

# rpm -q livecd-tools
livecd-tools-013-1.fc8

If not, install them as root:

yum install livecd-tools

Then run live-iso-to-disk as root:

# livecd-iso-to-disk Fedora-9-Beta-Live-i686.iso /dev/sdb1
Verifying image...
Fedora-9-Beta-Live-i686.iso: 1df4a4bc641220fa44cfd5d252dadb16
Fragment sums: 74221c82498ad77168fb66d935bd358ba8ca1727cbf3b5bf55259ff692c6
Fragment count: 20
Percent complete: 100.0% Fragment[20/20] -> OK
100.0
The media check is complete, the result is: PASS.

It is OK to use this media.
Copying live image to USB stick
Updating boot config file
Installing boot loader
USB stick set up as live image!

Note that your usb drive has to be bootable:

# fdisk -l /dev/sdb
Disk /dev/sdb: 1031 MB, 1031798272 bytes
255 heads, 63 sectors/track, 125 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x91f72d24

Device Boot Start End Blocks Id System
/dev/sdb1 * 1 126 1007584 6 FAT16
Partition 1 has different physical/logical endings:
phys=(124, 254, 63) logical=(125, 112, 50)

Leave your usb drive and reboot your pc on it.

bootchart - How fast are you booting?

2008-03-18T17:53:00.004+01:00

With bootchart, you are able to mesure how fast your pc is booting. You just need to install the package:
yum install bootchart
The following instruction (init=/sbin/bootchartd) is added in the /etc/grub.conf file in the kernel line.
Then reboot your pc. A data file is created in /var/log/bootchart.tgz. Then just run bootchart to generate a png file.

Git's user manual

2008-03-11T18:36:00.007+01:00

Here is a pdf version of the Git's user manual (v1.5.3).

problem with checkgmail 2

2008-02-28T09:24:00.004+01:00

Actually, I noticed that when I setup a check for a specified label that contains a '@' character (see label_delay in ~/.checkgmail/prefs.xml)

<label_delay mylabel="300"
me@domain.com="300">

then i get an error message when I rerun checkgmail.

not well-formed (invalid token) at line 15, column 25, byte 451 at /usr/lib/perl5/vendor_perl/5.8.8/i386-linux-thread-multi/XML/Parser.pm line 187
A thread exited while 2 threads were running.

The parser does not accept such character. Avoid them for now.

Problem running checkgmail

2008-02-27T11:37:00.003+01:00

I had the following problem when running checkgmail.

$ checkgmail
not well-formed (invalid token) at line 15, column 29, byte 495 at /usr/lib/perl5/vendor_perl/5.8.8/i386-linux-thread-multi/XML/Parser.pm line 187
A thread exited while 2 threads were running.

I just removed the .checkgmail directory in $HOME and re-ran it.

GNU/Linux distribution timeline

2007-11-28T12:38:00.001+01:00

I recently found an interesting image about the History of all the Linux distributions displayed in a timeline graphic.

(click on the image to see the original picture)

Linux is mainly 3 distributions: Debian, Slackware and Red Hat. They mainly differ by their packaging system (Red Hat: rpm, Debian: deb, Slackware: tgz).
It is unbelievable how hundreds of distributions have been created from them.

F8: Using iwl3945 instead of ipw3945 with Intel Corporation PRO/Wireless 3945ABG

2007-11-17T12:56:00.001+01:00

Until now, I was using ipw3945 from freshrpms.net. It was working fine but I decided to try using iwl3945 which is provided by the kernel itself.

Here is my network controller:

# /sbin/lspci | grep "Network controller"
0c:00.0 Network controller: Intel Corporation PRO/Wireless 3945ABG Network Connection (rev 02)

You need to install (if not already installed) the firmware corresponding to the kernel installed. I am using kernel-2.6.23.1-49.fc8 with the firmware iwl3945-firmware-2.14.1.5-2.

Check if the firmware is installed:
# rpm -qa iwl*
iwl3945-firmware-2.14.1.5-2

Check if the module is loaded by your kernel:
# /sbin/lsmod | grep 3945
iwl3945 159541 0
mac80211 112461 1 iwl3945

Check if your network controller is recognized with module iwl3945:
# dmesg | grep iwl3945
iwl3945: Intel(R) PRO/Wireless 3945ABG/BG Network Connection driver for Linux, 1.1.19kds
iwl3945: Copyright(c) 2003-2007 Intel Corporation
iwl3945: Detected Intel PRO/Wireless 3945ABG Network Connection
iwl3945: Tunable channels: 13 802.11bg, 23 802.11a channels

Add the following in /etc/modprobe.conf:
alias wlan0 iwl3945

The wlan0 interface is configured as follows:
# more /etc/sysconfig/network-scripts/ifcfg-wlan0
# Intel Corporation PRO/Wireless 3945ABG Network Connection
DEVICE=wlan0
BOOTPROTO=dhcp
HWADDR=00:18:de:b2:de:17
ONBOOT=no
DHCP_HOSTNAME=tosca.opera.univ-pau.fr
IPADDR=
DOMAIN=
NETMASK=
TYPE=Wireless
ESSID=
CHANNEL=1
MODE=Master
RATE=Auto
USERCTL=yes
IPV6INIT=no
PEERDNS=yes

The version of iwl3945 driver:
# more /sys/module/iwl3945/version
1.1.19kds

Install pam_keyring package. gnome's keyring (pam_gnome_keyring.so) does not work properly (it keeps asking me the passphrase of my wifi).

Do the following modifications in file /etc/pam.d/gdm

# more /etc/pam.d/gdm
#%PAM-1.0
auth [success=done ignore=ignore default=bad] pam_selinux_permit.so
auth required pam_env.so
auth optional pam_keyring.so try_first_pass ---- add this line
auth include system-auth
#auth optional pam_gnome_keyring.so ---- comment that one
account required pam_nologin.so
account include system-auth
password include system-auth
session required pam_selinux.so close
session include system-auth
session required pam_loginuid.so
session optional pam_console.so
session required pam_selinux.so open
session optional pam_keyinit.so force revoke
session required pam_namespace.so
#session optional pam_gnome_keyring.so auto_start ---- comment that one
session optional pam_keyring.so ---- add this line

and reboot. Ensure that NetworkManager will start during boot time.
chkconfig NetworkManager on

F8: Compiz-fusion does not work properly.

2007-11-14T17:29:00.001+01:00

I noticed when I installed gnome-compiz-manager that I did not have all the effects I had when I was running F7.
So i installed ccsm-0.6.0-3 and compiz-manager-0.6.0-3. Then add compiz-manager to the startup program list of Gnome. You will then be able to use all the features of compiz-fusion.

This is a known bug in F8.

F8: Yumex loops over and over (problem with file comps-f8.xml)

2007-11-11T23:23:00.001+01:00

If you get errors like the following

11:33:20 : Loading mirror speeds from cached hostfile
11:33:31 : Failure getting ftp://ftp.free.fr/mirrors/fedora.redhat.com/fedora/linux/updates/8/i386/repodata/comps-f8.xml:
11:33:31 : --> [Errno -1] Metadata file does not match checksum
11:33:31 : Trying other mirror.
11:33:43 : Failure getting ftp://ftp.ciril.fr/pub/linux/fedora/linux/updates/8/i386/repodata/comps-f8.xml:
11:33:43 : --> [Errno -1] Metadata file does not match checksum
11:33:43 : Trying other mirror.
11:33:53 : Failure getting ftp://ftp.proxad.net/mirrors/fedora.redhat.com/fedora/linux/updates/8/i386/repodata/comps-f8.xml :
11:33:53 : --> [Errno -1] Metadata file does not match checksum
11:33:53 : Trying other mirror.
11:34:07 : Failure getting ftp://ftp.cru.fr/pub/linux/fedora/updates/8/i386/repodata/comps-f8.xml:
11:34:07 : --> [Errno -1] Metadata file does not match checksum
11:34:07 : Trying other mirror.

there is certainly a problem with the sha1sum value of the file /var/cache/yum/updates/comps-f8.xml in the file /var/cache/yum/updates/repomd.xml, so replace the sha1sum value with the one of the file /var/cache/yum/updates/comps-f8.xml.

After modification, the file /var/cache/yum/updates/repomd.xml looks like this:

[...]
<data type="group">
<location href="repodata/comps-f8.xml">
<checksum type="sha">fea3dc6aef34f316fdf61d1e9ac2a61faae7e35c</checksum>
<timestamp>1194644692</timestamp>
</location>
</data>
[...]

F8: Errors when running compiz-fusion

2007-11-11T10:30:00.001+01:00

Here are some errors I had when running compiz in a terminal:

extension "XVideo" missing on display ":0.0"
compiz (core) - Fatal: No sync extension

To fix this, I added the followng line in the Module section:

Load "extmod"

and restarted the X server (Shift-Ctrl-Backspace).

F8: Live upgrade from F7 to F8

2007-11-10T12:49:00.001+01:00

Finally, Fedora 8 has been released yesterday. I decided this to do a live upgrade to F8.
Of course, it did not work 'out of the box'.

Here is what i did:

Install fedora-release, fedora-release-notes, livna-release with rpm -Uvf,
Disable all other repositories (in /etc/yum.repo.d, set enabled=0 in the repo files),
Start upgrading packages: yum --downloadonly upgrade,
I removed some packages that were causing some problems: lirc, vlc, vlc-plugin, xine-lib-moles, fuse, fuse-kmdl-2.6.23.1-21.fc7, ieee80211-kmdl-2.6.23.1-21.fc7 , ipw3945, ipw3945-kmdl-2.6.23.1-21.fc7, ntfs-3g , ntfs-config.
run again: yum --downloadonly upgrade,
After downloading everything, run yum upgrade,
It complained about gecko-libs: Error: Missing Dependency: gecko-libs = 1.8.1.8 is needed by package yelp, I solved the problem by installing devhelp (yum install devhelp),
It complained again about libsmbios:
Transaction Check Error:
file /usr/lib/libsmbios.so.1.0.13 from install of libsmbios-libs-0.13.10-1.fc8 conflicts with file from package libsmbios-0.13.12-1.fc7
file /usr/lib/libsmbiosxml.so.1.0.13 from install of libsmbios-libs-0.13.10-1.fc8 conflicts with file from package libsmbios-0.13.12-1.fc7
I had installed the package called libsmbios-0.13.12-1.fc7 which is more recent than the one of F8 libsmbios-libs-0.13.10-1.fc8 , so I removed the one of F7 (rpm -e -nodeps libsmbios-0.13.12-1.fc7),
run yum upgrade again, it is installing all the rpms downloaded previously,
reboot and cross fingers.
....
Yes! It works!

Increasing virtual disk space under VMware

2007-09-07T10:20:00.001+01:00

If you run out of space in your guest OS system running under VMware, you need to increase. If your physical partition on your Linux host is too small, you can increase it using gparted liveCD. It is similar to PartitionMagic but it is based on a gentoo, a linux distribution.
Once you have increased your partition on your Linux host,

you need to increase the virtual partition in VMware using the command:
vmware-vdiskmanager -X 50Gb -t 3 myos.vmdk
It will increase the space reserved for your OS up to 50 Gb and preallocate the space split in file of 2Gb size,
then use the iso image of the gparted liveCD and boot VMware on this image. It will launch gparted and you can modify the partitionning within VMware.

It seems that version 0.3.4-8 of gparted liveCD doesn't see the partitions of the guest OS. But it works well with version 0.3.4-5.

Wifi with Intel Corporation PRO/Wireless 3945ABG

2007-06-13T11:56:00.001+01:00

Wifi didn't work out of the box with Fedora 7. I had to install the following packages from atrpms:

ipw3945-1.2.0-18.3
ipw3945d-1.7.22-4
ipw3945-kmdl-2.6.21-1.3194
ipw3945-ucode-1.14.2-4
ieee80211-kmdl-2.6.21-1.3194

Just install them with

yum install ipw3945

and remove iwlwifi-firmware

yum remove iwlwifi-firmware

And then get the file http://svn.rpmforge.net/svn/trunk/rpms/ipw3945d/ipw3945d.init and copy it in /etc/init.d/ipw3945d and add the service:

chkconfig --add ipw3945d

and reboot.
You should see in the NetworkManager the wifi network along with the wired network.

F7: Dell Inspiron 9400 hangs on boot! (Part 2)

2007-06-13T08:30:00.001+01:00

I mentioned on June 1st that my Dell Inspiron 9400 was hanging on boot and that the solution was to add maxcpus=1 in /etc/grub.conf. If your laptop has 2 cpus, then only one is seen.
Another solution is to add clocksource=acpi_pm instead of maxcpus=1, like this:

title Fedora (2.6.21-1.3194.fc7) root (hd0,1) kernel /vmlinuz-2.6.21-1.3194.fc7 ro root=/dev/VolGroup01/LogVol00 rhgb quiet vga=792 clocksource=acpi_pm initrd /initrd-2.6.21-1.3194.fc7.img
and reboot.
Check in /proc/cpuinfo that you really have 2 cpus.

Setting up special Multimedia keys

2007-06-12T16:58:00.001+01:00

Edit the file /etc/X11/Xmodmap and add the following lines:
keycode 160 = XF86AudioMute keycode 174 = XF86AudioLowerVolume keycode 176 = XF86AudioRaiseVolume keycode 162 = XF86AudioPlay keycode 144 = XF86AudioPrev keycode 153 = XF86AudioNext keycode 164 = XF86AudioStop
Reboot and then run gnome-keybinding-properties and set up Volume mute, Volume Down, Volume Up, Play/Pause, Stop, Next track and Previous track.

Nvidia seems to work on Fedora 7

2007-06-08T12:53:00.001+01:00

I installed Fedora 7 recently (See hardware list here) and installed the nvidia driver from livna:

livna-release-7-2 (contains everything about livna repository. Install this one first.)
kmod-nvidia-1.0.9762-1.2.6.21_1.3194.fc7
xorg-x11-drv-nvidia-1.0.9762-2.lvn7
livna-config-display-0.0.12-1.lvn7

with the following kernel:

Linux 2.6.21-1.3194.fc7 #1 SMP Wed May 23 22:35:01 EDT 2007 i686 i686 i386 GNU/Linux

Got no problem (yet!).

Here is my /etc/X111/xorg.conf config file, the default one after installation:

# Xorg configuration created by livna-config-display

Section "ServerLayout"
   Identifier "Default Layout"
   Screen 0 "Screen0" 0 0
   InputDevice "Keyboard0" "CoreKeyboard"
   InputDevice "Synaptics" "CorePointer"
EndSection

Section "Files"
   ModulePath "/usr/lib/xorg/modules/extensions/nvidia"
   ModulePath "/usr/lib/xorg/modules"
EndSection

Section "InputDevice"
   Identifier "Keyboard0"
   Driver "kbd"
   Option "XkbModel" "pc105"
   Option "XkbLayout" "fr"
EndSection

Section "InputDevice"
   Identifier "Synaptics"
   Driver "synaptics"
   Option "Device" "/dev/input/mice"
   Option "Protocol" "auto-dev"
   Option "Emulate3Buttons" "yes"
EndSection

Section "Device"
   Identifier "Videocard0"
   Driver "nvidia"
   Option "AddARGBGLXVisuals" "True"
   Option "DisableGLXRootClipping" "True"
EndSection

Section "Screen"
   Identifier "Screen0"
   Device "Videocard0"
   DefaultDepth 24
   SubSection "Display"
     Viewport 0 0
     Depth 24
   EndSubSection
EndSection

I made some test before and after installing the nvidia driver and kernel module with glxgear:

before: 4782 frames in 5.1 seconds = 943.771 FPS
after : 64953 frames in 5.0 seconds = 12990.522 FPS

I installed it a few days ago and it did not work. Then I waited a few days and decided to reinstall it and it worked. Don't ask me what I did, I have no idea what so ever.

F7: Dell Inspiron 9400 hangs on boot!

2007-06-01T13:27:00.000+01:00

Fedora 7 was finally launched on May 31, 2007.
My Dell laptop (it has an Intel dual core CPU) did not want to boot correctly so I checked Fedora home page and found out that I had to add maxcpus=1 in the /etc/grub.conf file.

title Fedora (2.6.21-1.3194.fc7)
root (hd0,1)
kernel /vmlinuz-2.6.21-1.3194.fc7 ro root=/dev/VolGroup01/LogVol00 rhgb quiet maxcpus=1
initrd /initrd-2.6.21-1.3194.fc7.img

Check at fedora or at bugzilla.

Otherwise, installation went fine. I will now check more closely.

Fedora 7 to be launched

2007-05-16T08:24:00.000+01:00

Fedora 7 (and not Fedora Core 7) should be launched on Thursday May 24. Looking forward to install and test it.

"Fedora 7 has been about improving the manner in which all future Fedora releases will be made."
"[...] customized versions of Fedora are now possible to an extent that was not available previously."
"Kernel-based Virtual Machine (KVM) technology has been integrated with the Fedora graphical virtualization manager tool."

Problem running Adobe Reader 7.0

2007-04-06T12:50:00.001+01:00

Running acroread in a term won't open the adobe reader. And when opening a pdf file in Firefox, i get a message in a window (see picture).

The only way to fix this is to either:

put export GTK_IM_MODULE=scim-bridge in /usr/bin/acroread right after the line #!/bin/bash.
or to have export GTK_IM_MODULE=scim-bridge in your .bash_profile.

Logical Volume Manager

2007-02-02T11:11:00.001+01:00

Logical Volume Manager is a nice way to manage partitions on a computer (see here). It enables you to increase or decrease the size of Logical Volumes very easily.
The packages needed are lvm2 and system-config-lvm.
The only partition that must NOT be in a logical volume is /boot. Bootloaders do not understand LVMs yet. If you put /boot in an LVM, your system will not boot.
If you buy a disk (because you need space), you can easily add it to a Volume Group and create a new Logical Volume made of part or all of the Physical Volume.

This image shows the logical volumes I created. The LogVol03 is selected and it shows which physical volumes are composing the selected logical volume.

FC6 on Dell Inspiron 9400/E1705

2007-02-02T10:12:00.001+01:00

System

CPU	Intel Core Duo T7200 2GHz/4Gb cache
Graphic card	NVidia GeForce Go 7900 GS
Wireless	Intel PRO/Wireless 3945 802.11a/g Mini Card (54Mbps)
Screen	17'' UltraSharp TrueLife Wide-screen WUXGA (1900x1200)
Memory	2048 Mb (2X1024 ) 533MHz DDR2 SD
Hard disk	Samsung SATA 120Gb 5400 RPM
CD/DVD	8X DVD+/-RW Drive
Battery	80 WHr 9-cell Lithium Ion Primary Battery
Network	Broadcom BCM4401-B0 100Base-TX

I had absolutely no problem installing FC6 on my laptop. I used lvm2 to manage the different disk partitions. This enables to resize any partition very easily, especially when one of them becomes full.
The wireless works fine (see post of January 9).
I did not try the bluetooth feature (no device for that).

lspci output:
00:00.0 Host bridge: Intel Corporation Mobile 945GM/PM/GMS/940GML and 945GT Express Memory Controller Hub (rev 03)
00:01.0 PCI bridge: Intel Corporation Mobile 945GM/PM/GMS/940GML and 945GT Express PCI Express Root Port (rev 03)
00:1b.0 Audio device: Intel Corporation 82801G (ICH7 Family) High Definition Audio Controller (rev 01)
00:1c.0 PCI bridge: Intel Corporation 82801G (ICH7 Family) PCI Express Port 1 (rev 01)
00:1c.1 PCI bridge: Intel Corporation 82801G (ICH7 Family) PCI Express Port 2 (rev 01)
00:1c.3 PCI bridge: Intel Corporation 82801G (ICH7 Family) PCI Express Port 4 (rev 01)
00:1d.0 USB Controller: Intel Corporation 82801G (ICH7 Family) USB UHCI #1 (rev 01)
00:1d.1 USB Controller: Intel Corporation 82801G (ICH7 Family) USB UHCI #2 (rev 01)
00:1d.2 USB Controller: Intel Corporation 82801G (ICH7 Family) USB UHCI #3 (rev 01)
00:1d.3 USB Controller: Intel Corporation 82801G (ICH7 Family) USB UHCI #4 (rev 01)
00:1d.7 USB Controller: Intel Corporation 82801G (ICH7 Family) USB2 EHCI Controller (rev 01)
00:1e.0 PCI bridge: Intel Corporation 82801 Mobile PCI Bridge (rev e1)
00:1f.0 ISA bridge: Intel Corporation 82801GBM (ICH7-M) LPC Interface Bridge (rev 01)
00:1f.2 IDE interface: Intel Corporation 82801GBM/GHM (ICH7 Family) Serial ATA Storage Controller IDE (rev 01)
00:1f.3 SMBus: Intel Corporation 82801G (ICH7 Family) SMBus Controller (rev 01)
01:00.0 VGA compatible controller: nVidia Corporation GeForce Go 7900 GS (rev a1)
03:00.0 Ethernet controller: Broadcom Corporation BCM4401-B0 100Base-TX (rev 02)
03:01.0 FireWire (IEEE 1394): Ricoh Co Ltd Unknown device 0832
03:01.1 Class 0805: Ricoh Co Ltd R5C822 SD/SDIO/MMC/MS/MSPro Host Adapter (rev 19)
03:01.2 System peripheral: Ricoh Co Ltd Unknown device 0843 (rev 01)
03:01.3 System peripheral: Ricoh Co Ltd R5C592 Memory Stick Bus Host Adapter (rev 0a)
03:01.4 System peripheral: Ricoh Co Ltd xD-Picture Card Controller (rev 05)
0c:00.0 Network controller: Intel Corporation PRO/Wireless 3945ABG Network Connection (rev 02)

How I installed a Virtual Machine in VMware using a Ghost image.

2007-01-24T11:09:00.001+01:00

When I got new laptop, I decided to make a ghost image of all the partitions that were present on the disk.
It generated 4 files, among them C_Drive001.v2i and MYPC.sv2i (the name of my pc is MYPC). But VMware needs a vmware image (*.vmdk) so I converted the Ghost image C_Drive001.v2i to a vmware image (or virtual machine file) Windows XP Professional.vmdk using a tool available on vmware's website, VMware Virtual Machine Importer .

You give in input to the Importer the file MYPC.sv2i and choose Select volumes and resize to save or add space to make an vmware image only from the file drive C: otherwise you will get an vmware image of each partition. You don't want that. You can also specify the final size of your vmware image. A reasonable size would be 10Gb. You can anyway increase the size of your virual machine file later using vmware-vdiskmanager. Choose as destination type a VMware standalone virtual machine. Also you can choose Allocate all disk space now for better performance and you don't need to split files into 2Gb files. For the network part, choose NAT to share host's IP address.

Then I partitioned my disk and installed FC6. Later, I installed VMware 5.5.3, the Linux version on FC6. Linux is the Host OS. Windows XP Professional is the Guest OS.
Create a new VM to install the Windows XP Professional using the wizard. Go through all the steps and when you arrive at Select a disk, choose Use an existing virtual disk and select the Windows XP Professional.vmdk file obtained from the Importer tool.

You are done. Just power on the virtual machine, log in, install the VMware tools and activate your Windows XP Professional.

Reinstalling grub after having installed windows

2007-01-15T22:39:00.001+01:00

If you have a dual boot system Linux/windows and you need to reinstall windows (because you got a virus for example) or if you have Linux installed and decide to install windows, the installation destroys the MBR (Master Boot Record). So you need to reinstall grub.
The easiest way is to boot on the Linux rescue CD . When you get the prompt,

chroot /mnt/sysimage
grub-install /dev/sda

assuming that /dev/sda is the disk on which you installed Linux. You can check using the command fdisk as root:

fdisk -l

or as user

sudo /sbin/fdisk -l

to get:

Disk /dev/sda: 120.0 GB, 120034123776 bytes
255 heads, 63 sectors/track, 14593 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id System
/dev/sda1               1          10       80293+ de Dell Utility
/dev/sda2   *          11        2178    17414460    7 HPFS/NTFS
/dev/sda3            2179        4346    17414460    c W95 FAT32 (LBA)
/dev/sda4            4347       14593    82309027+   5 Extended
/dev/sda5            4347        4410      514048+ 83 Linux
/dev/sda6            4411       14593    81794916   8e Linux LVM

After installing grub, just type CTRL-D twice to reboot. You will then the grub menu.

That's all, Folks!

Environment modules, a way to dynamically modifiy user's environment

2007-01-10T12:16:00.001+01:00

Environment is a very nice tool to modify your environment dynamically using modulefiles.
Let's say you have two versions of ifort (the Intel compiler), 8.1.034 and 9.1.040. You would easily switch from one to the other without having to modify your .bash_profile file.
You need to install the package environment-modules-3.2.3-3.fc6. Then run the following command: /usr/share/Modules/bin/add.modules . this will add
in .bash_profile:

if [ -f /etc/profile.modules ]

then
. /etc/profile.modules
# put your own module loads here
module load null
fi

and in .bashrc

case "$0" in
          -sh|sh|*/sh) modules_shell=sh ;;
       -ksh|ksh|*/ksh) modules_shell=ksh ;;
       -zsh|zsh|*/zsh) modules_shell=zsh ;;
    -bash|bash|*/bash) modules_shell=bash ;;
esac
#module() { eval `/usr/share/Modules/$MODULE_VERSION/bin/modulecmd $modules_shell $*`; }
module() { eval `/usr/bin/modulecmd $modules_shell $*`; }

Modules should be installed in directory /usr/share/Modules/modulefiles. So for the ifort compiler, I would create in this directory another directory named ifort which will contain to files named respectively 8.1.034 and 9.1.040.
For example, file 9.1.040 will contain the following:

  1 #%Module 1.0
 2 #
 3 #  MPICH module for use with 'environment-modules' package:
  4 # 
 5 conflict ifort
 6
 7 set     version          9.1.040
 8
 9 prepend-path            PATH            /opt/intel/fc/$version/bin 
10 prepend-path            LD_LIBRARY_PATH /opt/intel/fc/$version/lib
11 prepend-path            MANPATH         /opt/intel/fc/$version/man 
12 setenv                  CFLAGS          -I/opt/intel/fc/$version/include
 13 setenv                  LDFLAGS         -L/opt/intel/fc/$version/lib 
14 prepend-path            INTEL_LICENSE_FILE      /opt/intel/licenses

The first line is an important magic cookie followed by a version number.
Line 5 says that you cannot load 2 versions of the same module.
Line 7 sets a variable named version local to the module which is called in lines 9 to 13 with $version (which is expanded to its value: 9.1.040).
Line 9 prepends the path where ifort is located to environment variable PATH. To append a path, use append-path instead.
Line 12 creates the environment variable CFLAGS using setenv (like tcsh does).

In the file /etc/profile.modules, you will specify the modules to be loaded at login time:

module load ifort/9.1.040

So when I login, which ifort should give me /opt/intel/fc/9.1.040/bin/ifort.
To switch to version 8.1.034, just do

module swap ifort/9.1.040 ifort/8.1.034

Here are other useful sub-command to command module:

module list
to list loaded modules
module avail
to list available modules
module unload ifort/8.1.034
to unload the module ifort version 8.1.034
module show ifort/8.1.034
to show the content of the modulefile ifort/8.1.034

The module /usr/share/Modules/modulefiles/use.own, if loaded at login time, creates a directory $HOME/privatemodules where each user can create their own modules.
He can then load them.

module initadd mymod
to add module mymod to the list of loaded module at login time.
module initrm mymod
to remove module mymod from the list of loaded module at login time.

Anyway, have a look at the man pages of module and modulefile.

Wifi with FC6

2007-01-09T10:08:00.000+01:00

I recently bought a DELL Inspiron 9400 with a PRO/Wireless 3945ABG Network Connection. I found the 802.11a/b/g driver ipw3945 in the Kwizart repository (see this post).
The packages to install are the following:

ipw3945
ipw3945d
ipw3945-ucode
ieee80211

With the NetworkManager, NetworkManager-glib and NetworkManager-gnome packages, it becomes very easy to connect to a wifi network ro to switch from a wired network to a wireless network.
More at the ipw3945 sourceforge page.

How to format a partition in ext3 format

2007-01-08T18:50:00.001+01:00

First, use fdisk to partition your disk using non-lvm format.
Then use the mke2fs command : mke2fs -j -L Label /dev/sda1

About yum and rpm commands

2007-01-08T13:16:00.000+01:00

Fedora distributions (as well as RedHat) used RPM packaging (Redhat Package Management). A user can install packages very easyly.

RPM
We must distinguish 3 kinds of packages: Binary packages (name.i386.rpm) and source packages (name.src.rpm) and architecture-independant packages (name.noarch.rpm). Binaries in the binary packages can be used right after installation. Source packages need to be compiled. Architecture-independant packages usually contains perl or bash or whatever scripts and thus don't need to be compiled.

To install (as root)
FC6 # rpm -ivh mypackage.rpm
or
FC6 # rpm -Uvh mypackage.rpm (if a version of that package is already installed)
To find a installed package on the system (can be done as user)
FC6 $ rpm -qa | grep kernel
This will find all installed packages whose name contains the word kernel
To uninstall (or remove) (as root)
FC6 # rpm -e kernel-2.6.18-1.2849.fc6 kernel-devel-2.6.18-1.2849.fc6
To list the content of a package (as user)

if installed on the system
FC6 $ rpm -ql kernel
if not installed
FC6 $ rpm -qpl mypackage.rpm

To list installed packages (as user)

sorted by size
FC6 $ rpm -qa --qf "%-10{SIZE} %-30{NAME}\n" | sort -n | less
sorted by name
FC6 $ rpm -qa --qf "%-30{NAME} %-10{SIZE}\n" | sort -n | less

To generate rpm packages

from a spec file and sources (tarball files like tar.gz or tar.bz2)
You need to install the package rpmdevtools-5.3-1.fc6 (as root) and then run as user fedora-buildrpmtree. This creates the directory $HOME/rpmbuild.
Copy your spec file in $HOME/rpmbuild/SPEC and your sources in $HOME/rpmbuild/SOURCES and run the following in the SPEC directory as user:
FC6 $ rpmbuild -ba myfile.spec
This will unpack your sources in $HOME/rpmbuild/BUILD, configure and compile the sources and generate binary rpms in $HOME/rpmbuild/RPMS/i386 and source rpms in $HOME/rpmbuild/SRPMS.
from a source rpm (as user):
FC6 $ rpmbuild --rebuild mypackage.src.rpm

To find which package a file belongs to (as user)
FC6 $ rpm -qvf /usr/bin/tree
tree-1.5.0-4
To rebuild the rpm database (as root)
If, for any reason, your rpm database seems corrupted, you can regenerate it using the following command:
FC6 # rpm --rebuilddb

YUM (Yellow dog Updated, Modified)
Yum works a bit differently from rpm in the sense that it downloads packages from repositories before installing them. Therefore your system must have information about the location of repositories which is located in files in directory /etc/yum.repo.d.

To update all packages installed (as root)
FC6 # yum update

To install (as root)
FC6 # yum install packagename
To find a package (as user)
FC6 $ yum list kernel
or with a wildcard (*)
FC6 $ yum list kernel*
To get info on a package (as user)
FC6 $ yum info kernel
To remove a package from the system (as root)
FC6 # yum remove kernel-2.6.18-1.2849.fc6 kernel-devel-2.6.18-1.2849.fc6
To run multiple commands of yum in a single yum session:
FC6 # yum shell
At prompt, you can run several commands such as list, info, install update, remove. The command ts shows what yum will perform.

Beside these 2 commands, there are several GUI that performs manipulations of packages such as pup or yumex.

My favorite repositories

2007-01-08T12:44:00.000+01:00

Here are my favorite yum repositories (to install these *.noarch.rpm, just run rpm -ivh *.noarch.rpm):

FreshRpms (http://ftp.freshrpms.net/pub/freshrpms/fedora/linux/6/freshrpms-release/freshrpms-release-1.1-1.fc.noarch.rpm)
Remi Collet (http://remi.collet.free.fr/rpms/fc6.i386/remi-release-1-2.fc6.remi.noarch.rpm)
kwizart (http://kwizart.free.fr/fedora/kwizart-release-6-1.noarch.rpm)
http://dries.ulyssis.org/rpm/
http://rpmforge.net/

File /etc/yum.repos.d/rpmforge.repo contains:
# Name: RPMforge RPM Repository for Fedora Core 6 - i386
# URL: http://rpmforge.net/
[rpmforge]
name = Fedora Core 6 - i386 - RPMforge.net - dries
#baseurl = http://apt.sw.be/dries/fedora/fc6/$basearch/dries/RPMS
#mirrorlist = http://apt.sw.be/dries/fedora/fc6/mirrors-rpmforge
mirrorlist = file:///etc/yum.repos.d/mirrors-rpmforge
enabled = 1
gpgkey = file:///etc/pki/rpm-gpg/RPM-GPG-KEY-rpmforge-dries
gpgcheck = 1

File /etc/yum.repos.d/mirrors-rpmforge contains:
http://apt.sw.be/dries/fedora/fc6/$ARCH/dries/RPMS
http://archive.cs.uu.nl/mirror/dag.wieers/dries/fedora/fc6/$ARCH/dries/RPMS
http://ftp2.lcpe.uni-sofia.bg/freshrpms/pub/dag/dries/fedora/fc6/$ARCH/dries/RPMS
#http://ftp.heanet.ie/pub/freshrpms/pub/dag/dries/fedora/fc6/$ARCH/dries/RPMS
http://ftp-stud.fht-esslingen.de/dag/dries/fedora/fc6/$ARCH/dries/RPMS
http://mirror.cpsc.ucalgary.ca/mirror/dag/dries/fedora/fc6/$ARCH/dries/RPMS
http://mirrors.ircam.fr/pub/dag/dries/fedora/fc6/$ARCH/dries/RPMS
http://rh-mirror.linux.iastate.edu/pub/dag/dries/fedora/fc6/$ARCH/dries/RPMS
http://rpmfind.net/linux/dag/dries/fedora/fc6/$ARCH/dries/RPMS
http://wftp.tu-chemnitz.de/pub/linux/dag/dries/fedora/fc6/$ARCH/dries/RPMS
http://www.mirrorservice.org/sites/apt.sw.be/dries/fedora/fc6/$ARCH/dries/RPMS

How to resize or expand LVM partitions

2007-01-08T10:25:00.000+01:00

A few week ago, I installed FC6 on my laptop and my slash (/) partition started to run out of space. So I decided that I would shrink my big partition and expand my / partition. Fortunately, I installed FC6 using LVM2 (Logical Volume Manager) which makes it very easy to resize partitions without even rebooting.

Here is how I did.
My partition was like this:

Filesystem	1K-blocks	Used	Available	Use%	Mounted on
/dev/mapper/VolGroup00-LogVol00	7935392	6773500	752292	91%	/
/dev/sda5	497829	20904	451223	5%	/boot
tmpfs	1037084	0	1037084	0%	/dev/shm
/dev/mapper/VolGroup00-LogVol01	70877776	14988144	51045372	23%	/home

My / partition is about 8 Gb and my /home partition is about 71 Gb.

For /home, I did:

umount /home
e2fsck -f /dev/VolGroup00/LogVol01
resize2fs /dev/VolGroup00/LogVol01 69G
lvreduce -L-2G /dev/VolGroup00/LogVol01
mount /home

For / partition, I did:

lvextend -L+2G /dev/VolGroup00/LogVol00
resize2fs /dev/VolGroup00/LogVol00

e2fsck and resize2fs belong to package e2fsprogs-1.39-7.fc6.

resize2fs has the ability to resize mounted partitions.
From the manual of resize2fs:

[...] If the filesystem is mounted, it can be used to expand the size of the mounted filesystem, assuming the kernel supports on-line resizing. (As of this writing, the Linux 2.6 kernel supports on-line resize for filesystems mounted using ext3 only.). [...]

[root@tosca ~]# resize2fs /dev/VolGroup00/LogVol00
resize2fs 1.39 (29-May-2006)
Filesystem at /dev/VolGroup00/LogVol00 is mounted on /; on-line resizing required
Performing an on-line resize of /dev/VolGroup00/LogVol00 to 2400256 (4k) blocks.
The filesystem on /dev/VolGroup00/LogVol00 is now 2400256 blocks long.

I finaly end up with the following partionning:

Filesystem	1K-blocks	Used	Available	Use%	Mounted on
/dev/mapper/VolGroup00-LogVol00	9299624	6779304	2043564	77%	/
/dev/sda5	497829	20904	451223	5%	/boot
tmpfs	1037084	0	1037084	0%	/dev/shm
/dev/mapper/VolGroup00-LogVol01	68877776	14999888	51033628	23%	/home

Some useful LVM commands (package lvm2-2.02.06-4):

lvmdiskscan
vgdisplay
pvdisplay
lvdispay
vgcreate
pvcreate
lvscan
pvscan
vgscan

Some useful links: