Downgrade PHP 5.3 to 5.2

Like many others, I recently upgraded my Production server to PHP 5.3. I’m not quite sure of the reason anymore. I think I was trying to integrate a PHP profiler that required something in 5.3. Whatever my original reason, after I got 5.3 installed on my server, it started slowly hemorrhaging memory and occasionally freezing up, always during the middle of the night while I slept of course. So, today I set about downgrading PHP 5.3.2 to PHP 5.2.13. Here is what worked for me.

The Setup

First off, I’m running Ubuntu-Hardy, and both my PHP 5.3 and 5.2 distributions came from dotdeb.org. I have a Production server and a Test server that I try to keep as similar as possible. The first thing I did was make a backup of my Test server with PHP 5.3 installed and my app functional. Then I began hacking away on my Test server trying to successfully downgrade to PHP 5.2. When I screwed the Test server up beyond repair, I restored from backup. Likewise, when I made progress I documented my steps, restored from backup, and tried to repeat my success. I did this several times and eventually made a script that successfully downgraded PHP. I rehearsed this script 3 times on my Test server, backed-up my Production server, then repeated the script on my Production server. IT WORKED!

NOTICE: I probably have a couple of unnecessary steps in here that I didn’t bother to investigate their necessity. You may also need to uninstall and reinstall different PHP modules than I did, depending on the requirements of your applications.

My Final Steps

  1. Remove PHP
    1. sudo apt-get remove php5-common
    2. sudo apt-get remove php5-cli
    3. sudo apt-get remove php5
    4. sudo apt-get autoremove memcached
  2. Update your sources list to point to PHP 5.2
    1. sudo nano /etc/apt/sources.list
    2. Remove the references to PHP 5.3 packages. For me that was:
      deb http://php53.dotdeb.org stable all
      deb-src http://php53.dotdeb.org stable all
    3. Add the following 5.2 packages:
      deb http://http.us.debian.org/debian stable all
      deb http://security.debian.org/ stable/updates main contrib
      deb http://packages.dotdeb.org/ stable all
  3. Install PHP 5.2
    1. sudo apt-get update
    2. sudo apt-get install php5-cli
    3. sudo apt-get install php5
  4. Reinstall any PHP modules your application requires. For me that was:
    1. sudo apt-get install memcached
    2. sudo apt-get install php5-memcache
    3. sudo apt-get install php5-curl
    4. sudo apt-get install php5-mysql
  5. sudo reboot

After my production server rebooted, I had to fidget with Apache to get it started again. You may have to:

  • sudo /etc/init.d/apache2 restart
  • sudo /etc/init.d/apache2 stop
  • sudo /etc/init.d/apache2 start

Hope this helps!

Twitter MySQL User Object Table Schema

A common practice for Twitter application developers is creating a MySQL table to store user objects returned from Twitter’s API. I’m trying to optimize my user table for Friend Or Follow, and it occurred to me how many developers must be reinventing this MySQL table schema, so let’s get it right once and for all.

Here’s where I’m starting. If you see anything that I’m missing, or notice a data type that could be optimized let me know in the comments and I’ll update the post to reflect your suggestions.

Installing Cassandra on Ubuntu Linux

Cassandra is a highly scalable, eventually consistent, distributed, structured key-value store. It’s an alternative to SQL if you don’t need relational data structures. The best part is it’s super fast, and distributed. The worst part is there’s not much documentation or community, at least compared to MySQL. So, I’m doing my small part to contribute to Cassandra. Here’s how I installed Cassandra on Ubuntu 8.04.2 LTS (hardy), but these directions should work on pretty much any Linux distro.

1. First upgrade your software as is with the following two commands (just for good measure):

sudo apt-get update
sudo apt-get upgrade

2. Now, open up your Debian package sources list with Nano for editing using the following command:

sudo nano /etc/apt/sources.list

3. Next, add the following sources to your /etc/apt/sources.list file.

deb http://www.apache.org/dist/incubator/cassandra/debian unstable main
deb-src http://www.apache.org/dist/incubator/cassandra/debian unstable main

After you add these two lines, press cntrl+X to close Nano. It’ll ask “Save modified buffer?” Press Y. Press Enter when Nano asks “File Name to Write.”

4. Run the update to install Casandra with this command:

sudo apt-get update

5. ERROR! At this point you receive an error similar to this:

W: GPG error: http://www.apache.org unstable Release: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY F758CE318D77295D

6. Use the following three commands to ignore the signature error, and continue installing:

NOTE: You must replace the key value ‘F758CE318D77295D’ with the key value you received in your error message.

gpg --keyserver wwwkeys.eu.pgp.net --recv-keys F758CE318D77295D
sudo apt-key add ~/.gnupg/pubring.gpg
sudo apt-get update

7. Install Cassandra:

sudo apt-get install cassandra

8. Next you need to change Cassandra’s default port number from 8080 to something else, because the 8080 port typically conflicts with SSH terminal connections. Use Nano to open up the Cassandra configuration file using the following command:

sudo nano /usr/share/cassandra/cassandra.in.sh

9. Then change the port number 8080 on the following line to 10036, and save the file:

-Dcom.sun.management.jmxremote.port=10036 \

10. Start Cassandra with the command:

/etc/init.d/cassandra start

Once you have Cassandra running, test it with Cassandra’s command line tool CLI. Use the example found on the Cassandra Wiki.

How to Update Ubuntu Hardy to PHP 5.2.9

First upgrade your software as is.

>> sudo apt-get update

>> sudo apt-get upgrade

Then add the following to your /etc/apt/sources.list file.

>> sudo nano /etc/apt/sources.list

deb http://http.us.debian.org/debian stable all
deb http://security.debian.org/ stable/updates main contrib
deb http://packages.dotdeb.org/ stable all

Finally run the update to install PHP 5.2.9.

>> sudo apt-get update

>> sudo apt-get install php5-cli

>> sudo /etc/init.d/apache2 restart

Global page navigation accross WordPress MU blogs

Here’s the scenario. You have a WordPress MU site with multiple blogs, but for whatever reason, you want every blog to have the same main navigation and pages. Your main site/blog hosts all of your pages. So how do you go about creating a WordPress MU Theme that will use the navigation and pages from your main site/blog accross all of your blogs? Well, I’m here to tell you how I did it.

The WordPress Codex has a function called ‘wp_list_pages.’ What I wanted to do is create a new function called ‘wp_list_main_pages.’ So I simply searched my WordPress install and copied the ‘get_pages’ and ‘wp_list_pages’ functions and pasted them into my theme’s ‘functions.php’ file. From here I made a few edits to these functions. To start with, I renamed them ‘wp_list_main_pages’ and ‘get_main_pages.’

Below is the final code I’m using to display the main navigation accross blogs. I made notes in bold to help. Hope this is helpful to you Googling coders. 🙂

// You can use this function instead of 'wp_list_pages' in your theme
function wp_list_main_pages($args = '') {
	$defaults = array(
		'depth' => 0, 'show_date' => '',
		'date_format' => get_option('date_format'),
		'child_of' => 0, 'exclude' => '',
		'title_li' => __('Pages'), 'echo' => 1,
		'authors' => '', 'sort_column' => 'menu_order, post_title'
	);

	$r = wp_parse_args( $args, $defaults );
	extract( $r, EXTR_SKIP );

	$output = '';
	$current_page = 0;

	// sanitize, mostly to keep spaces out
	$r['exclude'] = preg_replace('[^0-9,]', '', $r['exclude']);

	// Allow plugins to filter an array of excluded pages
	$r['exclude'] = implode(',', apply_filters('wp_list_pages_excludes', explode(',', $r['exclude'])));

	// Query pages.
	$r['hierarchical'] = 0;
	// Right here we call our new 'get_main_pages' function
	$pages = get_main_pages($r);

	if ( !empty($pages) ) {
		if ( $r['title_li'] )
			$output .= '<li class="pagenav">' . $r['title_li'] . '<ul>';

		global $wp_query;
		if ( is_page() || $wp_query->is_posts_page )
			$current_page = $wp_query->get_queried_object_id();
		$output .= walk_page_tree($pages, $r['depth'], $current_page, $r);

		if ( $r['title_li'] )
			$output .= '</ul></li>';
	}

	$output = apply_filters('wp_list_pages', $output);

	// This line is sloppy and needs improvement.
	// You have to remove the name of the blog your currently on from you global navigation.
	// I'm doing this the simplest but least scalable way here.
	$output = str_replace("pressroom/", "", $output);

	if ( $r['echo'] )
		echo $output;
	else
		return $output;
}

// This is essentially a private function
function &get_main_pages($args = '') {
	global $wpdb;

	// This is the magic line.
	// Now when the SQL runs to pull your navigation pages, it'll use your main blogs ID.
	$wpdb->set_blog_id(1);

	// Notice here I call the original get_pages function and return the results
	$pages = get_pages($args);
	return $pages;
}