Time for a tech post. I had the privilege of upgrading an ancient Red Hat 9 Linux box to CentOS 4.5. Red Hat Linux has been EOL for a very long time, and was superceded by Fedora, which has had seven major releases. That’s the equivalent of running Windows 3.1 Well, not quite Anyhow, the box at hand appeared to be a custom web server. I will outline the steps I used to successfully upgrade it to CentOS 4.5.

1. Download the CentOS 4.5 CD set or DVD image from www.centos.org.
2. Put in CD 1 or the DVD image.
3. Reboot the old Red Hat 9 box and ensure that the BIOS allows booting from the CD/DVD drive.
4. Enter the following at the linux boot prompt and press Enter: linux upgradeany
5. The system will happily hum along while Red Hat 9 is upgraded to CentOS 4.5.
6. If ‘yum’ wasn’t installed, you will need to install it:
rpm -ivh http://centos.arcticnetwork.ca/4.5/os/i386/CentOS/RPMS/python-elementtree-1.2.6-5.el4.centos.i386.rpm
rpm -ivh http://centos.arcticnetwork.ca/4.5/os/i386/CentOS/RPMS/sqlite-3.3.6-2.i386.rpm
rpm -ivh http://centos.arcticnetwork.ca/4.5/os/i386/CentOS/RPMS/python-sqlite-1.1.7-1.2.1.i386.rpm
rpm -ivh http://centos.arcticnetwork.ca/4.5/os/i386/CentOS/RPMS/python-urlgrabber-2.9.8-2.noarch.rpm
rpm -ivh http://centos.arcticnetwork.ca/4.5/os/i386/CentOS/RPMS/yum-2.4.3-3.el4.centos.noarch.rpm
7. Now, run the following command and press Enter: yum update
8. Allow yum to run and update the required packages. Reboot if yum downloads and installs a new kernel.
9. Since the box in question had a custom Apache and PHP configuration, I reinstalled them manually:
yum install httpd
yum install php
yum install mysql
yum install mysql-server
yum install php-mysql
10. I had to update the MySQL configuration to point to the correct socket:
Update my.cnf:
socket=/var/lib/mysql/mysql.sock
instead of
socket=/tmp/mysql.sock
and add the following to use your existing database passwords
old_passwords=1
11. Copy any required PHP extensions to new system path:
cp -R /usr/local/lib/php/extensions/* /usr/lib/php4
12. Edit /etc/php.ini:
error_reporting = E_ALL & ~E_NOTICE
extension=mysql.so
13. Finally, set services to start up automatically:
chkconfig –level 345 httpd on
chkconfig –level 345 mysqld on
chkconfig –level 345 sendmail on
14. There you have it. The server should be happily running CentOS 4.5 now. Remember to keep it up-to-date by running yum update regularly.

Best of all, I upgraded the entire OS remotely using SSH. Upgrading your OS remotely is usually not a good idea, but if you know what you’re doing, go for it! Ensure you backup any custom/modified configuration files just in case you press a wrong key during the upgrade! The upgrade worked well, with the exception of recompiling eAccelerator (to work with PHP 5.2.2) and updating a line (see below) of code in WordPress to make everything work smoothly.

Immediately following the upgrade, I noticed that my WordPress blog was loading very slowly and I also noticed random strings of hexadecimal characters at the top of the main blog page.

If you upgrade your server to Ubuntu Server 7.04 and run WordPress 2.1.x, ensure that you change line 832 in wp-includes/functions.php to the following (all on one line):

@header($_SERVER["SERVER_PROTOCOL"].” “.$header.” “.$text, true, $header);

Once you make the above change, your blog should load normally.

Update (5.15.2007):

With WordPress 2.2 (released today), the aforementioned WordPress fix is now irrelevant.

Well, here it is!

By default, Ubuntu Server installs MySQL and PHP. However, it doesn’t tweak the system for performance. If you run a WordPress blog on your server, you may not realize that your blog could run more efficiently. WordPress relies on both PHP and MySQL. PHP scripts need to be read from a server’s hard drive, parsed by the web server software and then executed. Wouldn’t it be nice if the server could just skip the first two steps and just execute PHP scripts as quickly as plain HTML?

That’s where eAccelerator comes in. eAccelerator is a free, open-source PHP accelerator and cache. I’ve been using it for nearly seven years on personal and commercial servers. In other words, it’s very stable for a free product.

Let’s get started.

First of all, you need to ensure that you have the php5-dev package installed on your server. It’s likely not installed, so let’s install it:

you@yourserver:~$ sudo apt-get install php5-dev

Next, download the latest build and extract it:

you@yourserver:~$ wget http://bart.eaccelerator.net/source/0.9.5.1/eaccelerator-0.9.5.1.tar.bz2
you@yourserver:~$ tar -xvjf eaccelerator-0.9.5.1.tar.bz2
you@yourserver:~$ cd eaccelerator-0.9.5.1

Time to compile eAcccelerator:

you@yourserver:~$ sudo find / -name phpize
/usr/bin/phpize
you@yourserver:~$ export PHP_PREFIX=”/usr/bin”
you@yourserver:~$ $PHP_PREFIX/phpize
you@yourserver:~$ ./configure –enable-eaccelerator=shared –with-php-config=$PHP_PREFIX/php-config
you@yourserver:~$ sudo apt-get install make
you@yourserver:~$ make

That was a mouthful, eh? Next, we need to install the eAccelerator module we just created:

you@yourserver:~$ cd modules
you@yourserver:~$ sudo cp eaccelerator.so /usr/lib/php5/20051025/

Edit your php.ini and paste the following block into it after the extensions block:

zend_extension=”/usr/lib/php5/20051025/eaccelerator.so”
eaccelerator.shm_size=”4″
eaccelerator.cache_dir=”/tmp/eaccelerator”
eaccelerator.enable=”1″
eaccelerator.optimizer=”1″
eaccelerator.check_mtime=”1″
eaccelerator.debug=”0″
eaccelerator.filter=”"
eaccelerator.shm_max=”0″
eaccelerator.shm_ttl=”0″
eaccelerator.shm_prune_period=”0″
eaccelerator.shm_only=”1″
eaccelerator.compress=”1″
eaccelerator.compress_level=”7″

Almost done. Create /tmp/eaccelerator:

you@yourserver:~$ mkdir /tmp/eaccelerator
you@yourserver:~$ chmod 777 /tmp/eaccelerator

One more step! We need to restart Apache for the changes to take effect:

you@yourserver:~$ sudo /etc/init.d/apache2 restart

There you have it. Your PHP-based scripts and applications should run nearly 70% faster!

In the Ubuntu 6.10 / Edgy release notes, Ubuntu recommends the following procedure to upgrade from Dapper to Edgy:

# apt-get dist-upgrade && apt-get dist-upgrade

That will not work. Why? Because sources.list needs to be updated first (and apt-get update needs to be run beforehand).

Prepare to update sources.list by issuing the following command:

# sudo vi /etc/apt/sources.list

Replace all instances of dapper with edgy. My updated sources.list looked as follows:

deb http://us.archive.ubuntu.com/ubuntu/ edgy main restricted
deb-src http://us.archive.ubuntu.com/ubuntu/ edgy main restricted

deb http://us.archive.ubuntu.com/ubuntu/ edgy-updates main restricted
deb-src http://us.archive.ubuntu.com/ubuntu/ edgy-updates main restricted

deb http://security.ubuntu.com/ubuntu edgy-security main restricted
deb-src http://security.ubuntu.com/ubuntu edgy-security main restricted

Once you save the file, run the following string of commands:

# sudo apt-get update && sudo apt-get dist-upgrade && sudo apt-get dist-upgrade

Go for coffee. Run around the block. Play a game. You get the picture. It will take awhile to download all of the required packages (255 in my case).

Now, run the following commands to ensure everything was installed correctly.? The last line deletes the downloaded files from apt’s cache, freeing up disk space.

# sudo apt-get -f install
# sudo dpkg –configure -a
# sudo apt-get clean

Finally, cross your fingers and run the following command:

# sudo reboot

Hopefully, everything worked and your Ubuntu server is running the latest release!

If asked to overwrite a modified configuration file, reply ‘N’. Otherwise, you might end up with a broken system. In my case, I also had to recompile eAccelerator 0.9.5-rc1 as it was compiled for the version of PHP 5 (5.1.2) included with Ubuntu Server 6.06.

If you installed from a CD-ROM and have a working network connection to your Ubuntu system, you should update the sources.list file used by apt-get so that your system won’t prompt you to insert the Ubuntu CD-ROM each time you install a new package:

1. Log in to your system via the console.
2. Run the following command: sudo vi /etc/apt/sources.list
3. Add a # before the following line (use cursor keys to put cursor in front of the line and then press i, followed by #) :

deb cdrom:[Ubuntu-Server 6.06.1 _Dapper … restricted

4. Save sources.list by pressing ‘Esc’, followed by the following sequence of characters to write the file and exit: :wq

Next, update your system with the latest Ubuntu packages. Having the latest packages ensures that your system remains secure.

Run sudo apt-get update, followed by sudo apt-get upgrade. It may take awhile to download updated packages for your installation, depending on your Internet connection.

If you want to admin your server remotely, or via another box on your home network, type the following to install SSH: sudo apt-get install ssh.

Ensure that the date on your server is correct by running date:

you@yourserver:~$ date
Wed Sep 20 14:12:54 PDT 2006

Correct time zone, wrong time. If the time zone is incorrect, you will need to run the tzconfig utility.

Check to see if the system is set to use UTC or not.

you@yourserver:~$ cat /etc/default/rcS | grep “UTC”
# Set UTC=yes if your system clock is set to UTC (GMT), and UTC=no if not.
UTC=yes

Since you probably want the server to return the time in your local time zone, change the line to read: UTC=no

One more step. You’ll probably want to synchronize the system clock using an official source, provided that your time zone was set during installation. If so, simply execute the following:

you@yourserver:~$ sudo apt-get install ntp-simple ntp-server

Ensure that your firewall permits communication over UDP port 123 and you’re set!

Check back in a couple of days for some more tips.

I shelled in only to see one error message upon restarting the Apache web service: Segmentation fault. I checked the last few lines of Apache’s error_log only to see many more errors related to the same problem. It’s one of most dreaded messages one can receive on a Linux box, aside from the dreaded ‘kernel panic’.

A segmentation fault usually means one of two things in the Linux world: a software fault or a hardware fault. So, I set out to determine what the underlying cause was.

# apt-get update
Segmentation Fault

As a system package installation tool, apt-get usually runs without any errors (unless imposed by network issues). Perhaps, a corrupt filesystem was the culprit? I created a file in the root of the server, forcefsck, to force a file system scan on the next reboot. However, it turns out that it wasn’t such a good idea issuing the reboot command remotely, as the system was waiting for my input once I got home. It turns out there were multiple, soft disk errors waiting to be corrected. After correcting the file system errors, I ran MemTest86+ and also Seagate’s SeaTools to test the RAM and hard drive, respectively. Both turned out to be fine, with no errors. After some more software and hardware testing, the culprit was determined to be the motherboard. Argh! Over 10 GB of data needed to be backed up.

Four hours later (started at 7 PM, finished at 11 PM), the server was up and running on entirely new hardware (well, except for the hard drive). I decided to use Ubuntu Server 6.06.1 LTS for the server operating system this time around. After a couple of minor obstacles (eg. compiling and installing eAccelerator 0.9.5-rc1), everything is running smoothly again