CZ:Technical/How to set up a CZ clone on CentOS 5
From Citizendium, the Citizens' Compendium
This page provides instructions describing how to set up a Citizendium clone on the CentOS 5 operating system. Creating a CZ clone is useful for a number of purposes, including:
- Storing a local copy of Citizendium on a lap-top or other device so articles are accessible when there is no available network connection.
- Creating a local copy of Citizendium for the purpose of fixing bugs in the modified version of Mediawiki that forms its software base or developing extensions to that code.
Installing the necessary software infrastructure
CZ is implemented by a modified version of the Mediawiki software. This software relies on several software components that underly its operation, specifically:
- PHP - an object-oriented interpreted language. Mediawiki is written in PHP and its software modules execute when accessed through a web server.
- Apache2 - an open source industrial grade web server.
- Postgres - an open source industrial grade database system.
Together, these are frequently referred to as a LAPP stack (Linux Apache2 Postgres PHP stack). Note that many Mediawiki implementations use the MySQL database system, rather than Postgres. While it might be possible to install Citizendium's wiki content using a MySQL database, doing so would represent experimental development. This how-to presumes the installer uses Postgres.
There are a number of different platforms that support a LAPP stack and potentially any of them could be used to support a CZ clone. While these instructions describe how to set up a CZ clone on the CentOS 5 operating system, it is possible to install CZ's software on other Linux platforms and even on other operating systems (e.g., Mac OS X, Windows). There are instructions for setting up a CZ clone on Ubuntu. Those interested in using using an OS platform other than Ubuntu or CentOS 5 will have to do the research necessary to accomplish that.
It must be noted that unlike Ubuntu and other Debian Linux variants, CentOS is much more expert friendly and much less novice friendly. For those without significant experience with Linux systems (specifically, installing and administering them), the recommended OS for a CZ clone is Ubuntu. However, the production version of the CZ wiki runs on Fedora, the same base used for CentOS, so for those with the requisite experience and skills, implementing a CZ clone for development work on CentOS has certain advantages. But with those advantages comes risks. The yum software installation system on CentOS is prone to significant configuration problems, especially when setting up its repositories. So, those choosing to use CentOS do so at their own risk. Caveat Emptor.
These instructions have been tested on a fresh CentOS 5.4 install.
Setting up the CentOS 5.4 System
Bringing your CentOS 5.4 installation software up to date
Before installing your CZ clone, it is important to bring your CentOS 5.4 software up to date. Generally, soon after you boot your system, if there are software updates available, a notice window will pop-up informing you of this. However, this may not happen immediately for a fresh install. In that case, you can force a software update by selecting Applications at the top of your desktop, then select System Tools and finally Software Updater. This will start the sequence that queries the existing software repositories to see if your software is up to date. If not, you will be asked whether you wish to do so. You should answer in the affirmative and update your software before moving on to the next steps.
Turn off SELinux
By default CentOS 5.4 is installed with SELinux enabled. SELinux is security technology developed so CentOS would conform to certain criteria set by the U.S. government. This technology is not particularly user or system administrator friendly and frequently interferes with the proper execution of application software. In order to make these instructions simple and unencumbered with significant side trips into SELinux configuration, they assume SELinux is disabled. Setting up a CZ clone for a SELinux enabled system is beyond the scope of this article.
To disable SELinux, you must edit /etc/selinux/config, change a variable and then reboot your system. Open a terminal window and execute:
cd /etc/selinux sudo gedit config
Whent the edit window appears, change the line SELINUX=enforcing to SELINUX=disabled, save the file and exit the editor. After editing the configuration file should look like:
# This file controls the state of SELinux on the system. # SELINUX= can take one of these three values: # enforcing - SELinux security policy is enforced. # permissive - SELinux prints warnings instead of enforcing. # disabled - SELinux is fully disabled. SELINUX=disabled # SELINUXTYPE= type of policy in use. Possible values are: # targeted - Only targeted network daemons are protected. # strict - Full SELinux protection. SELINUXTYPE=targeted # SETLOCALDEFS= Check local definition changes SETLOCALDEFS=0
Now reboot your system and continue with the instructions in the next section.
The version of postgres we want is not available through the normal CentOS software repositories. So, we have to modify the repository information to direct yum to install the correct version. (Note: these instructions are based on those given at Yet Another Guide.
First find out what is the latest install package for postgres 8.3 (install packages for CentOS have the suffix rpm). Using a new tab or browser window, follow the link postgres rpms. Three files should display in the browser window. One of these will have the name pgdg-centos-8.3-x.noarch.rpm, where x will be a number, such as 6. Now download the appropriate yum repository configuration by entering the following at a terminal command prompt (replacing the x in the file name with the integer just discovered):
cd /tmp sudo wget http://yum.pgsqlrpms.org/reporpms/8.3/pgdg-centos-8.3-x.noarch.rpm
Then install this information using the following command (again changing the x to the appropriate integer):
sudo chmod +x pgdg-centos-8.3-x.noarch.rpm sudo rpm -ivf pgdg-centos-8.3-x.noarch.rpm
If you are working on a CentOS installation on which some other software has already been installed, it is possible that other repositories are indicated in the yum configuration files that may interfere with the installation of postgres 8.3. You can determine this as folows. Execute the following commands:
cd /etc/yum.repos.d sudo gedit CentOS-base.repo &
If this file is non-empty, then do the following. (If the file is empty, simply close it without making any changes). There are sections in the configuration file, each headed by a word in square brackets. If there are two sections labeled [base] and [updates] then at the bottom of those sections add the line:
Click on the Save button at the top of the edit window and close it.
Now execute the following commands:
sudo yum install postgresql sudo yum install postgresql-server
For each command a bunch of text is displayed followed by a line that specifies the total download size. The next line is a prompt: Is this ok [y/N]:. You must type y. The default is N, which (of course) means no and taking it will abort the install.
The install of postgres on CentOS does not initialize the directory that postgres uses to store database information. So, the next step is to do this. Execute the following commands:
sudo /etc/rc.d/init.d/postgresql initdb
Installing postgres doesn't mean the server starts up when the system boots. We have to configure the system to do that. First, check that chkconfig is in your execution path. At a command prompt enter:
If the error bash: chkconfig: command not found is returned, you will have to add /sbin to $PATH. This requires editing .bash_profile (remember to cd to your home directory):
cd ~ gedit .bash_profile &
In the edit window there should be a line starting with $PATH. At the end of this line add :/sbin (don't forget the colon at the beginning of this text and there should be no space between the colon and the rest of the line). Then save the edit, exit the editor and type:
source .bash_profile chkconfig
This should result in an error message about usage. When chkconfig is working, enter the following commands:
sudo chkconfig --add postgresql sudo chkconfig postgresql on
Now reboot your system and when that completes open a terminal window. Type:
su su postgres psql
A welcome message should display, followed by hints on psql commands and then the prompt postgres=#. At this prompt type \q.
Your terminal identity was changed from your CentOS username to root and then to postgres. Get back to root by typing:
You should see a # prompt. There is one more thing we have to do before proceeding. In CentOS, the default trust model is ident based. This causes problems when configuring the postgres databases using pgAdmin III. So, we need to make one change to the postgres configuration file. Execute the following commands at a terminal window prompt:
cd /var/lib/pgsql/data gedit pg_hba.conf &
An edit window should appear. Scroll to the end of the file and find the entry with the comment "local" is for Unix domain socket connections only. The next entries control access when connections to the postgres server come from the local host. Copy the text below and replace the corresponding text. The entry should now look like the following:
# "local" is for Unix domain socket connections only #local all all ident sameuser local all all trust # IPv4 local connections: #host all all 127.0.0.1/32 ident sameuser host all all 127.0.0.1/32 trust # IPv6 local connections: #host all all ::1/128 md5 host all all ::1/128 trust
The change comprising the last three lines is only necessary if the system is using IPv6 addressing.
Save the file and exit the editor. Now restart the postgres server with the following command:
This change allows you to manipulate postgres through pgAdmin III without changing your CentOS user identity using su or sudo. The install of postgres is now complete. You can exit from root or remain at root and ignore the sudo prefixes in the next section.
Installing apache2 and PHP
If you are still logged in as root in your terminal window, you can either execute the following commands without the sudo or exit root and return to your normal user identity.
Installing postgres is a lot of work. Fortunately, installing the other two components of the LAPP stack takes much less effort. First, we will install the apache2 web server. In a terminal window execute the following command:
sudo yum install httpd
Follow the normal install procedure by answering y to the question. You may be prompted with a question whether it is OK to import a GPG key. If so, answer y.
After the install completes, execute the following two commands:
sudo chkconfig --levels 235 httpd on sudo /etc/init.d/httpd start
Then in your favorite browser, create a new tab or window and in the URL field type: http://localhost. An apache2 test page should display. If so, apache2 is installed. If not, then make sure you correctly executed all of the commands specified above.
To install PHP5 is even simpler. Execute the following commands:
sudo yum install php sudo yum install php-pgsql
If you want to test that PHP5 is working properly, execute:
cd /var/www/html sudo gedit phpinfo.php &
When the empty edit window appears, enter the following information:
<?php phpinfo(); ?>
Save the edit and then execute:
sudo /etc/init.d/httpd restart
This command restarts the apache2 server so it is aware that PHP5 is available. Now in a browser create a new tab or window and in the URL field type:
A page containing information about the PHP5 installation should appear, followed by information about installed PHP5 extensions.
Installing the Subversion client
In order to install the CZ software, the installer must use the subversion client. CentOS 5.4 doesn't come with subversion , so it must be manually installed. At a terminal prompt execute:
sudo yum install mod_dav_svn subversion
After this installation completes, the LAPP stack and other software for creating a CZ clone is available.
Configuring the LAPP stack
Once the LAPP stack is installed, the installer must configure it.
The first thing to do is assign a password to the postgres user of the postgres database. This step may be somewhat confusing to a novice user of postgres, so some explanation is in order. On the operating system that runs postgres is a user named postgres. This user is also listed as a Login Role for the postgres database system. The next step uses a feature of postgres that stipulates if you access the database (through the utility psql) as a particular user, you are automatically logged into postgres as that user. So, the first step is to use sudo to login as root and then change the user identity with the -u parameter. The command executed by sudo will use psql to connect to the postgres database (which, while having the same name, is different than the postgres Login Role). Once psql is connected to the postgres database using the Login Role postgres, it is possible to change the password for (the Login Role) postgres. Execute the following command at a terminal prompt:
sudo -u postgres psql postgres
This should result in a welcome message, followed by some command hints and then followed by the prompt:
At this prompt type:
psql will prompt you for a password and then prompt you to reenter it (remember this password, because you will need it later). If you successfully enter the same password twice, psql returns a new prompt. Type "\q" (without the quotes), which should return you to a Ubuntu command prompt.
The above instructions use the term postgres in three different ways. There is the postgres database system, which is the software that implements postgres. There is the postgres database, which is a collection of tables managed by the postgres database system. Finally, there is the postgres Login Role, which is an identify used by the postgres database system to make access control decisions. It is necessary to keep these distinctions in mind. Later in these configuration instructions, we willl create a cz Login Role and a cz database.
For the novice, using postgres can be a frustrating experience. The main command line utility for interacting with postgres is "psql". Psql has two sets of commands. The first always begin with the character '\' (the backslash, not the slash). The second are SQL commands. Frequently, when using the second category of commands, a new user fails to enter the required semicolon (";") at the end. (The semicolon is not used at the end of the first category of commands). After entering an SQL command, psql returns a prompt and the user thinks the command has been executed. However, it hasn't. A semicolon must be entered (it can be entered after the command on the subsequent prompt). To make things even more confusing, psql uses a "query buffer". This concatenates all input until the requisite semicolon is typed. This also makes things difficult for the novice user, since many times he/she will type multiple commands and then remember to type one with a semicolon. This almost always results in an error. It is good practice to clean out the query buffer using the command "\r" (without the quotes) before entering a new SQL command. It is recommended that those unfamiliar with psql read this good tutorial on its use.
A GUI for accessing postgress databases, pgAdmin III, can help the novice user avoid many of the frustrations experienced when using psql. To install pgAdmin III, use Ubuntu's Add/Remove application facility. This is accessed by clicking at the top left of the desktop on the "Applications" tag (Postgres Figure 1). Select All available applications in the Show drop down menu and type postgres in the Search box (Postgres Figure 2). When the search finishes, scroll down to the bottom and select the pgAdmin III application by checking the box to the left. Then click on Apply Changes. This will install pgAdmin III on your system. (The system may request you to enable community applications. If so, click "Enable"). You will also have to enter your superuser password in a prompt window.
Postgres Figure 1
Postgres Figure 2
Once installed, pgAdmin III requires configuration. Start pgAdmin III (it should appear in the Applications menu under Programming). The application window appears. In the left window pane is an icon labeled Servers(0). It is necessary to connect to a postgres server in order for the tool to operate. Under the File menu, select Add Server. A New Server Registration window pops up (Postgres Figure 3).
We are going to use the tool to connect to the database server on the local system. So, fill in the Name field with an appropriate name (e.g., Local Host) and fill in the Host field with the value localhost. The Username field is filled in by default to postgres. We will use this identity to access the database system until we have other Login Roles established. So, it is necessary to fill in the password created for the postgres Login Identity previously. Once this is filled in (leave all other selections at their default values), the OK button becomes active. Click it. You may see a hint displayed. If you find these annoying, there is a check box that turns them off. The result of this is a new entry under Servers(1) appears called Local Host (note the number of servers has changed from 0 to 1). Click on the + sign to the left of Local Host and a number of other labels appear.
We will use pgAdmin III to create two new Login Roles. The first is a role corresponding to your user name. This will allow you to use psql from your user account without specifying a username and password. Click on Local Host so it is highlighed and then click on the Edit menu and select New Object. A pop out menu appears to the right. Select New Login Role.... Fill in the form as follows (Postgres Figure 4a/4b):
- Role name - your user name on the system. For example if you login to Ubuntu as jdoe, fill in jdoe into the Role name field.
- Password - enter a password for your postgres Login Role (this is a postgres specific password. Normally you would not use your Ubuntu password.).
- Password (again) - enter the same password.
- Role Privileges - click all of the boxes. This gives you all privileges for database administration.
Click on OK.
Postgres Figure 3
Postgres Figure 4a
Postgres Figure 4b
To check that you successfully created a Login Role with the same username as your CentOS name, type the following command at a terminal window prompt (don't close pgAdmin III. We will use it again below):
This should result in the postgres welcome message, command hints and then the prompt: postgres=>. Type \q to exit psql (if you wish, you can stay in psql rather than quitting, since we will return to it after some more work in pgAdmin III).
Going back to pgAdmin III, we will create another Login Role as well as a new database. To create the Login Role follow the steps given above, except use the Role Name cz and leave the password fields blank. While it is unnecessary to give the cz Login Role all privileges, this clone of CZ will be used by only you and it is therefore convenient to do so. If you were creating an installation of CZ that would be accessed by others, you would spend much more time configuring it so it is secure. How to configure a CZ installation so it is secure in a multi-user environment is outside the scope of these instructions.
We have to modify the cz role and using psql is the easiest way to do so. In a terminal window at a prompt type:
This should result in the postgres welcome message, command hints and then the prompt: postgres=#. Now enter the following information at the postgres prompt. (note: you can simply copy and past the information below into the terminal window).
ALTER ROLE cz SET "TimeZone" TO 'GMT'; ALTER ROLE cz SET search_path TO mediawiki, public; ALTER ROLE cz SET "DateStyle" TO ISO, YMD; ALTER ROLE cz SET client_min_messages TO 'error';
If you enter this information one line at a time, you will have to type a return each time (don't forget the semicolon after each line). If you copy and paste the information, you will need to type a return after the last line. As each line is executed, psql will return ALTER ROLE. To check that these commands executed correctly, use pgAdmin III to navigate to the cz Login Role (i.e, expand the + sign next to Login Roles). Double click on cz. A pop-up window should appear with the information for cz. Select the Variables tab and the variable values set by psql should display (Postgres Figure 5). You may need to refresh the information displayed by pgAdmin III by selecting Refresh from the View menu. You can exit psql by typing "\q".
Now that a cz Login Role is established with the correct initialization information, it is possible to create a cz database. In the main pgAdmin III window click on Databases so it is highlighted and either right click on it and select New Database or from the Edit Menu select New Database (in some versions of pgAdmin III New Database is a selection in the New Object submenu). Fill in the Name field with cz. Select cz using the right hand down-arrow of the Owner field (Postgres Figure 6). Click OK. The Databases field should change from Databases(1) to Databases(2). Click on the + sign to the left of the Databases field and two databases should be displayed: 1) postgres, and 2) cz (Postgres Figure 7).
Postgres Figure 5
Postgres Figure 6
Postgres Figure 7
This completes the configuration of postgres. We will return to postgres when we create the CZ clone database.
Some additional configuration of the apache2 server is necessary to utilize the CZ software that implements the wiki. There are two ways to configure the apache2 server for the CZ clone. One way is to assume there is only one copy of the CZ software on the machine. This is normally true for those who are cloning CZ for personal use. The other way is to assume multiple copies of the CZ software exist on the machine. This is normally the case if the CZ clone is used for software development. While setting up the machine for multiple copies of the CZ software is unnecessary for a local use clone, it really doesn't introduce any overhead to do so. Consequently, these instructions describe how to set up the apache2 server to access more than one copy of the CZ software.
The CZ software will be stored in subdirectories of /usr/local/src. The first step is to cd to that directory and create a subdirectory called mediawiki. Change the permissions on mediawiki to 777 (i.e., in /usr/local/src execute sudo chmod 777 mediawiki). In a production environment, the adminstrator would use more careful access control permissions. However, since this CZ clone is intended for local use, making the software world r/w is an acceptable configuration strategy. Change directories to mediawiki and create a CZ_1_13_2 subdirectory. Also make this directory world r/w. The commands are:
cd /usr/local/src sudo mkdir mediawiki sudo chmod 777 mediawiki cd mediawiki sudo mkdir CZ_1_13_2 sudo chmod 777 CZ_1_13_2
If the CZ clone is used for development purposes, the developer would create additional subdirectories of the /usr/local/src/mediawiki directory to hold different versions of the CZ software. In fact, it is possible to host copies of the non-modified mediawiki software by downloading them into subdirectories of mediawiki. For example, if it is desirable to download a copy of the mediawiki software on which CZ is based, the developer could create a subdirectory of mediawiki, say MW_1_13_2, and use it to hold the non-modified version.
Once this directory structure is created, it is possible to make the changes to the apache2 configuration information necessary to support the CZ software. In a terminal window exectute:
cd /etc/httpd/conf sudo gedit httpd.conf
This will open an edit window displaying the contents of the http configuration file. At the end of this file enter the following information:
Alias /mediawiki/ "/usr/local/src/mediawiki/" <Directory "/usr/local/src/mediawiki/"> Options FollowSymLinks AllowOverride None Order allow,deny Allow from all </Directory> Alias /CZ_1_13_2 "/usr/local/src/mediawiki/CZ_1_13_2/phase3/index.php"
Save the file, close the editor and restart the apache2 server by executing:
sudo /etc/init.d/httpd restart
We can't test whether the apache2 server properly executes the CZ software because that software isn't installed yet. However, we can test the configuration changes we just made. In a terminal window execute:
su cd /usr/local/src/mediawiki/CZ_1_13_2 mkdir phase3 chmod 777 phase3 cd phase3 cp /var/www/html/phpinfo.php . mv phpinfo.php index.php chmod 777 index.php exit
The first command makes you root (after entering the root password) so you do not have to type sudo before commands. The last command, exit, takes the session back to your CentOS identity.
These commands make a copy of the phpinfo.php file in /usr/local/src/mediawiki/CZ_1_13_2/phase3 and rename it index.php. When we install the CZ software we will have to remove this file before loading it, since there is already an index.php file in the CZ software distribution.
Now we are prepared to test the new apache2 server configuration. Open a browser and either open a new tab or a new browser window. In the URL field enter:
The same PHP configuration information should appear as before. Now enter in the URL field:
Again the PHP configuration information should appear. If either of these tests fail, recheck the edits and the directory structure created above. In the latter case, the directories /usr/local/src/mediawiki, /usr/local/src/mediawiki/CZ_1_13_2 and /usr/local/src/mediawiki/CZ_1_13_2/phase3 should exist with 777 permissions.
Installing the CZ wiki software
Download the CZ wiki software
Installing the CZ wiki software requires access to the main CZ repository. In this repository is stored various versions of the software including the currently deployed version and others undergoing development. Before downloading this software, we have to clean up some detritus left from testing the apache2 configuration. Execute the following commands in a terminal window:
cd /usr/local/src/mediawiki/CZ_1_13_2 sudo rm -fr phase3
This removes the phase3 directory and everything in it. Now we have to find out which release version we will checkout. Execute:
svn list http://test.citizendium.org/czrepo/tags
The result is a list of releases for the CZ software. These instructions assume the latest version is called CZ_1_13_2_1.
The next step is to checkout this version. For those new to version control, a checkout from a repository makes a local copy of whatever branch is specified. This local copy retains information about where it comes from in the repository, but the repository itself keeps no state about the local copy. Consequently, you can do anything you want with the local copy, even delete it. The repository is unaware of these changes until you attempt to check in the local copy.
To checkout CZ_1_13_2_1 into the directory CZ_1_13_2 (which is where will configure the apache2 server to find it), execute:
cd /usr/local/src/mediawiki svn checkout http://test.citizendium.org/czrepo/tags/CZ_1_13_2_1 CZ_1_13_2
The result of this command is a very large number of lines starting with the letter A (which stands for Add). After the command completes, there is a copy of the latest CZ released software in /usr/src/local/mediawiki/CZ_1_13_2.
Setting up the CZ database
The next step is to create the database schema. In an unmodified version of the mediawiki software the schema is created using a utility called update.php. However, CZ runs a modified version of the mediawiki software and so this utility doesn't correctly setup the database. So, we must use psql to import the schema.
The directory /usr/src/local/mediawiki/CZ_1_13_2 has two subdirectories: 1) database, and 2) phase3. The latter contains the PHP modules that implement the wiki. The former contains information specific to the wiki database. In particular, it contains a dump of the CZ database schema. We will use psql to import this into the cz database.
In a terminal window execute:
cd /usr/local/src/mediawiki/CZ_1_13_2/database psql -d cz -U cz -f schema_1_13_2.pg
If an error occurs with the following text:
FATAL: Ident authentication failed for user cz
then you haven't correctly changed the file pg_hba.conf. Go back and ensure that both local and host entries are specified as trust.
Before configuring the CZ wiki software, we will make one modification to the CZ database. It isn't clear why this change is necessary, but if it isn't executed, the database import will not work. Furthermore, this action creates a dummy user with a user id of 0. A real user cannot have this user id, since the software will not allow such a user to login.
Start pgAdmin III. Double click the localhost icon and if they are not already expanded, click on the + button to the left of the cz icon and then on the + button to the left of the Schemas icon and then on the + button to the left of the mediawiki icon. Under mediawiki click on the + button to the left of the Tables icon. Scroll down to the mwuser entry. Single click mwuser and then click on the icon that looks like a table at the top of the window shown in Postgres Figure 8.
Postgres Figure 8
Postgres Figure 9
In the first row under user_id, enter the integer 0. In the next column, user_name, enter the text "dummy" (without the quotes). Select File in the upper left-hand corner and then select save. The result should look like Postgres Figure 9. These actions create a dummy user that cannot be used to login to the wiki, since it has no password associated with it.
In order to adjust the wiki state for the dummy user, we must now make a change to a database sequence variable. This category of database information provides consecutive values that aid in labeling other database data. Click the - to the left of Tables and then click the + to the left of Sequences. Scroll down to user_user_id_seq and double click it. Change current value from 0 to 1 and click OK. Close pgAdmin III.
Configuring the CZ wiki software
Most configuration of the CZ wiki software occurs through edits to the file LocalSettings.php. In a non-modified version of the Mediawiki (MW) software, after an initial software install using Subversion, the installer access the wiki through a web browser and a special PHP routine guides him/her through the steps necessary to achieve an intial configuration. However, since the CZ wiki uses a different database schema than MW, the CZ software distribution comes with a copy of LocalSettings.php that is already properly configured. It is only necessary to get a copy of LocalSettings.php into the phase3 directory of the CZ software distribution. In a terminal window execute:
cd /usr/local/src/mediawiki/CZ_1_13_2/phase3 cp config/LocalSettings.php .
Now, open a browser window and in the URL field type:
The wiki main page should appear. If so, congratulations! You have successfully installed the CZ wiki software. If you wish, you can bookmark this URL and save it under a more informative title, such as "CZ clone" or "Local CZ clone". If the main page did not appear, go back and review the instructions in this section to ensure you executed them correctly.
Establishing the first user
User registration in a CZ wiki requires an administrator to approve them. This is to ensure they are using the real names and they provide a short description of themselves that goes on their user page. However, when the software is first installed, no users exist (ignoring the dummy user, which has no login capability). So, the first user is established using a one-off procedure. This involves using a utility in the maintenance directory of the wiki software. This utility installs a user with specified privileges.
In a terminal window execute:
cd /usr/local/src/mediawiki/CZ_1_13_2/phase3/maintenance php createAndPromote.php --bureaucrat Wikiadmin Wikiadminpw
For "Wikiadminpw" you can substitute any password you wish. If the createAndPromote call exits with the error:
cz: Creating and promoting User:Wikiadmin...PHP Warning: pg_query(): Query failed: ERROR: duplicate key value violates unique constraint "mwuser_pkey" in /usr/local/src/mediawiki/CZ_1_13_2/phase3/includes/db/DatabasePostgres.php on line 552
you forgot to change the current value of user_user_id_seq to 1. If you execute the call again it should work.
Post wiki software installation and configuration
While all of the software necessary to create a CZ clone is described in the previous sections, there is other software installation and configuration required to support user registration and to support optional, but commonly provided features. First we must get local email working on the machine and configure the CZ software to use it so user registration works. Then we must install and configure the texvc software in order for mathematical equations to display.
Configuring local email
In order to logon to your CZ clone, your account must have a valid email address configured. The way the CZ software judges that this condition obtains is to send you a confirmation email when you first register. This applies to all users, including the Wikiadmin user created using createAndPromote.php. So, in order to continue, we must first configure email. Normally, email is configured by a system admin to allow users to send and receive messages from other users on the internet. If your system already is configured to do this, nothing else is required. Skip the next sub-section and proceed to Setting up your email address. Otherwise, you must complete the steps in the next section.
Setting up local host email
Setting up local host email requires the installation of software and then some configuration activity.
Installing the required software
We need to install two software applications to enable local host email: 1) Postfix, a Mail Transport Agent, and 2) the Mail Transport Agent Switcher.
To install Postfix, access the Applications->Add/Remove Software menu and click the Search tab. In the search field enter Postfix and click the search button. A number of options should appear. Select the one with the comment Postfix Mail Transport Agent and click the Apply button.
Now see if the mail transport agent switcher is installed. If you cannot find this application in the System->Administration menu, search for it in the Add/Remove Software submenu under Applications. Add both the switcher and the switcher gui.
Configuring local host email
Setting up local email on a system is not hard. Since there is a good set of instructions that describe how to do this elsewhere, they are not duplicated here. The installer should proceed to How to set up email in CentOS 5. There are two things to note about those instructions, however. Firstly, the instructions tell you to open the file browser and double click the file main.cf in the directory /etc/postfix. However, this doesn't work, since you need root privileges to edit the file. So, instead in terminal window execute the following commands:
cd /etc/postfix sudo gedit main.cf
Secondly, since we are setting up local email, there is no need to remove the comments from the two lines:
#inet_interfaces = all #mydestination = $myhostname, localhost.$mydomain, localhost, $mydomain
The configuration file is already set up for local email. The only change required is to uncomment the line:
#home_mailbox = Maildir/
so it reads:
home_mailbox = .Maildir/
Don't forget to put a period before Maildir/ (i.e., it should be .Maildir/ not Maildir/). This means the mail directory will not appear in a vanilla ls listing.
Setting up the email client
Once local email is enabled, we must set up an email agent to read the local email spool file. (If you already can send and receive email on your system, skip this section and proceed to Installing and configuring texvx.
From the Applications menu at the top left corner of the Desktop, select Internet and then Email. This displays the Evolution email client configuration window.
From the Welcome page click Forward and Forward again. The Identity page should display. In the window type <your username on the local system>@localhost. For example, if your username on the system is jdoe, then enter jdoe@localhost. Click the Forward button. The next window should be labeled Receiving Email. From the pop-down menu labeled Server Type select Maildir-format mail directories (this assumes you followed the instructions in the previous section and selected Maildir formated email). From the Configuration pop-down menu select Other... and when the Mailbox location window appears, type ~/.Maildir in the Location: field. Click Forward. In the next (Receiving Options) window, select how often you want Evolution to check for email. Then click Forward. In the Sending Email window choose Sendmail from the pop-down menu. Then click Forward. In the Account Management window select the name that you want to appear for this account in Evolution. Then click Forward. In the Timezone window select your time zone. Then click Forward and Apply.
Evolution should launch and in the left-hand column will be the name you chose in the Account Management window. You should be able to send and receive email from this account.
Configuring the CZ clone with your email address
If you have used the CZ wiki at Citizendium, you should already know how to configure your email address. However, for completeness this section describes that task.
Access your local CZ clone wiki (http://localhost/CZ_1_13_2) using a browser and login. At this point there is only one user - Wikiadmin, so login with this identity. When login completes, click on the Preferences tab at the top of the browser window. When your preferences window displays, enter your email address in the User profile field and click Save. A message informing you that a confirmation email has been sent appears. If Evolution is not running, start it (if it is running, click Send/Receive at the top left-hand corner). A confirmation email should appear in your account inbox. Follow its instructions. This will confirm your email address.
Installing and configuring texvx
Installing the required software
TeX support on the mediawiki software (on which the CZ software is based) requires the installation and configuration of several software packages. While for some this may go smoothly, others have had significant problems getting math equations to appear properly on wiki pages. Since this functionality is pretty basic to much of the content on CZ, there really is no option but to support it. However, be prepared for some frustration in getting this to work properly.
The software we need is not available on the software repositories that ship with CentOS 5.4. So, we need to add one. Note: third party yum software repositories can be hazardous to use. The original author of this article once added several third party repositories of questionable quality and after updating the system software, yum stopped working (yum is a python script and the update corrupted the python subsystem). So, you should be very careful when adding repositories to your yum configuration information.
In this case, we will add RPMForge. This third party repository has a good reputation and so should not be dangerous to use. However, there are no guarantees and since those installing a CZ clone on CentOS 5.4 should be experienced CentOS installers, they are expected to have the skills to get out of problems caused when a third party repository corrupts their system.
The first thing to do is to install yum priorities. In a terminal window execute:
sudo yum install yum-priorities
As always, when prompted whether you wish to continue, answer y.
Make sure priorities are enabled (they are in a fresh CentOS 5.4 install) by executing:
This should produce the following output:
[main] enabled = 1
If enabled is not set to 1, edit the file and change it so it is.
Now, cd to /etc/yum.repos.d and edit each file by setting the priority for each repository listed. The recommended settings (given by the CentOS wiki at http://wiki.centos.org/AdditionalResources/Repositories/RPMForge) are:
[base], [addons], [updates], [extras] ... priority=1 [centosplus],[contrib] ... priority=2 Third Party Repos such as rpmforge ... priority=N (where N is > 10 and based on your preference)
For example, execute the following in a terminial window:
cd /etc/yum.repos.d sudo gedit CentOS-Base.repo
When the edit window opens, you will see sections labeled by the section identifiers given above. For example:
[base] name=CentOS-$releasever - Base mirrorlist=http://mirrorlist.centos.org/?release=$releasever&arch=$basearch&repo=os #baseurl=http://mirror.centos.org/centos/$releasever/os/$basearch/ gpgcheck=1 gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-5 #released updates [updates] name=CentOS-$releasever - Updates mirrorlist=http://mirrorlist.centos.org/?release=$releasever&arch=$basearch&repo=updates #baseurl=http://mirror.centos.org/centos/$releasever/updates/$basearch/ gpgcheck=1 gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-5
After the gpgkey line add priority=1. These two sections should now look like:
[base] name=CentOS-$releasever - Base mirrorlist=http://mirrorlist.centos.org/?release=$releasever&arch=$basearch&repo=os #baseurl=http://mirror.centos.org/centos/$releasever/os/$basearch/ gpgcheck=1 gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-5 priority=1 #released updates [updates] name=CentOS-$releasever - Updates mirrorlist=http://mirrorlist.centos.org/?release=$releasever&arch=$basearch&repo=updates #baseurl=http://mirror.centos.org/centos/$releasever/updates/$basearch/ gpgcheck=1 gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-5 priority=1
Make sure you set the priorities correctly for all of the sections.
For the [centosplus],[contrib] sections, ensure the variable enabled is set to 1. That is:
Set the priority variable on sections in all other .repo files in the directory. The repository file CentOS-Media.repo has one repository c5-media. In a fresh install, this repository is not enabled. If your system has enabled it, set its priority to 3. Since we have loaded postgres, there may also be a repo file with the name pgdg-83-centos.repo. If so, set the pgdg83 entry to priority=11.
Now determine what is the machine architecture:
This should return either i386 or x86_64. Depending on this value in a terminal window execute:
sudo wget http://apt.sw.be/redhat/el5/en/i386/RPMS.dag/rpmforge-release-0.3.6-1.el5.rf.i386.rpm
sudo wget http://apt.sw.be/redhat/el5/en/x86_64/RPMS.dag/rpmforge-release-0.3.6-1.el5.rf.x86_64.rpm
rpm packages are signed, so in order to use them we have to download the key for RPMForge. In a terminal window execute:
sudo rpm --import http://dag.wieers.com/rpm/packages/RPM-GPG-KEY.dag.txt
Then verify the package:
sudo rpm -K rpmforge-release-0.3.6-1.el5.rf.*.rpm
Now install the repository into the systems repository list:
sudo rpm -i rpmforge-release-0.3.6-1.el5.rf.*.rpm
Check that the installation completed properly:
RPMForge should appear in the list of repositories and there should be activity showing the information associated with it is updated.
Now that RPMForge is added to the repository list, we must set its priority correctly. In a terminal window execute:
cd /etc/yum.repos.d sudo gedit rpmforge.repo
At the end of the [rpmforge] section add:
Save the file and exit the editor.
Texvc depends on other software on the system. Particularly, ImageMagick, Ghostscript, latex, dvipng and dvips must be installed. ImageMagick and Ghostscript are installed by default in CentOS 5.4. To ensure this, execute:
whereis ImageMagick whereis ghostscript
(make sure to capitalize the I and M in ImageMagick). This should result in the following output:
ImageMagick: /usr/share/man/man1/ImageMagick.1.gz ghostscript: /usr/bin/ghostscript /etc/ghostscript /usr/lib/ghostscript /usr/share/ghostscript /usr/share/man/man1/ghostscript.1.gz
or something similar.
If one or the other of these utilities is missing, you will have to install them with yum, i.e.,
sudo yum install ImageMagick
sudo yum install ghostscript
Normally latex and dvips are not installed on a standard CentOS 5.4 system. This is easily checked with the following commands:
whereis latex whereis dvipng whereis dvips
Generally either all will be available or none. To install all of them, in a terminal window execute:
sudo yum install tetex tetex-fonts tetex-dvips tetex-latex
After ensuring the availablity of ImageMagick, Ghostscript, latex, dvipng and dvips we must install ocaml. In a terminal window execute:
sudo yum install ocaml
Ocaml requires the availability of the gcc compiler. Check whether it exists by executing in a terminal window:
If this returns:
gcc: no input files
gcc is installed. Otherwise, execute:
sudo yum install gcc gcc-c++ autoconf automake
and answer y at the appropriate time.
We are finally in a position to build texvc. In a terminal window execute:
cd /usr/local/src/mediawiki/CZ_1_13_2/phase3/math make
Configuring texvc is simple, but non-intuitive. The reason is texvc is written in Ocaml and it calls various utility functions. In order for these to work correctly, certain environmental variables, permissions and directories must be properly set-up. This is the first step in getting texvc to work. In a terminal window execute:
cd /usr/local/src/mediawiki/CZ_1_13_2/phase3 chmod -R 777 images
These commands set the permissions on the images directory and all its subdirectories to world read-write-execute. For a production wiki installation this is probably not recommended, but since this is your personal wiki, it shouldn't be a problem.
Next, edit LocalSettings.php:
cd /usr/local/src/mediawiki/CZ_1_13_2/phase3 gedit LocalSettings.php
Change $wgTexvc to equal $IP/math/texvc. That is:
We now have to set up some directories. Execute:
cd /var/www sudo mkdir .texmf-var sudo chgrp apache .texmf-var sudo chmod 775 .texmf-var cd .texmf-var sudo mkdir web2c sudo chgrp apache web2c sudo chmod 775 web2c
Finally, there is a slight mismatch between CentOS 5.4 and texvc that we must accommodate. Specifically, the apache2 web server is launched by root at system boot. Consequently, it inherits root's HOME environmental variable. Regrettably, latex (called by texvc) uses HOME when looking for some format information. To correctly direct latex to the proper directory, we must modify the HOME environmental variable that is passed to apache when it starts. Fortunately, there is a simple way to do this. In a terminal window, execute:
cd /etc/sysconfig sudo gedit httpd
An edit window appears. At the end of the file add the following line (don't put spaces between the characters):
Save the file and close the edit window. We now have to restart the apache2 server. Execute:
sudo /etc/init.d/httpd restart
The server should stop and then start. To test texvc, open a browser window in your CZ clone (http://localhost/CZ_1_13_2) and create a page to edit (e.g., create a sandbox subpage for the Wikiadmin user). Edit the page and enter:
Click Show preview. The result should be:
If the equation above appears, then texvc is working properly. However, you may get one of several errors:
Failed to parse (Cannot write to or create math temp directory):
This occurs because the math/ and tmp/ directories in phase3/images are not accessible to the apache2 server. Check to ensure you have executed the first set of commands correctly.
Another common error is:
Failed to parse (Missing texvc executable; please see math/README to configure.):
If so, make sure you have correctly modified LocalSettings.php and have saved the results.
A pernicious error is:
Failed to parse (PNG conversion failed; check for correct installation of latex, dvips, gs, and convert)
This means the CZ wiki software is having trouble calling all of the installed software necessary to render the <math> markup. Make sure you have executed the third and fourth set of commands properly.
The four command sets given above properly configure a fresh CentOS 5.4 install so texcv works. If you are installing texvc on a system that is not a fresh install, you may have to work through other configuration problems. A good source of advice for doing that is found on the Mediawiki site. Good luck!
Loading the CZ database
The preceding sections describe how to load and configure the software required to run a CZ clone. After completing this, it is necessary to load the CZ database into the clone. This populates the clone with CZ page content. Note, however, that the public content from CZ contains neither history information nor image data. So, the database loaded into the CZ clone will only contain page text content.
Loading the clone database from a CZ daily dump file
The official way to load a content database into a Mediawiki software installation uses a utility in phase3/maintenance. (Note: Before following these instructions, read the next section Loading the clone database from a postgres database dump file. Doing so may save you a great deal of time.) The first step is to obtain the data dump file. This is available at CZ:Downloads. Two files are available, which are identical, except for how they are compressed. These instructions assume the reader uses the bzip2 version. The changes necessary to use the gzip version should be obvious.
Navigate to CZ:Downloads and click on the bz2 file link. A window will appear asking you where you wish to save the file. Either save it in your home directory or save it somewhere else and move/copy it to your home directory. Then, in a terminal window, execute:
cd ~ bunzip2 cz.dump.current.xml.bz2
This command decompresses the file and changes its name to cz.dump.current.xml. Once the .xml file is available in your home directory, in a terminal window execute:
cd /usr/local/src/mediawiki/CZ_1_13_2/phase3/maintenance php importDump.php ~/cz.dump.current.xml
Processing of the data dump file commences and importDump outputs incremental status lines. Output will look something like:
100 (1.57 pages/sec 1.57 revs/sec) 200 (1.49 pages/sec 1.49 revs/sec)
The integer at the beginning of each line specifies the number of pages importData has processed so far. When importDump completes, the clone database is available for use.
Between lines specifying the number of pages importData has processed, you may get output that looks like the following:
This is dvips(k) 5.96.1 Copyright 2007 Radical Eye Software (www.radicaleye.com) ' TeX output 2010.02.02:1219' -> </usr/share/texmf-texlive/dvips/base/tex.pro> </usr/share/texmf-texlive/dvips/base/texps.pro>. </usr/share/texmf-texlive/fonts/type1/bluesky/cm/cmsy10.pfb> </usr/share/texmf-texlive/fonts/type1/bluesky/cm/cmr12.pfb> </usr/share/texmf-texlive/fonts/type1/bluesky/cm/cmmi12.pfb>
This is normal.
After importing the data dump file, various internal database data structures will be out of date. The importDump operation creates jobs to correct this. However, normally a job is run per page access. If there are a large number of jobs (which is normally the case - sometimes in the tens of thousands), the normal page accessing activity on a clone will not provide sufficient page accesses to clear out the job queue. So, the clone maintainer must force the execution of these jobs. This is done with the maintenance script runJobs.php. Since this utility runs each job serially, running only one instance of it will still require a great deal of time to clear the job queue. Fortunately, runJobs is coded so it can be run concurrently. The following shell script runs 20 instances of runJobs:
#!/bin/bash cd /usr/local/src/mediawiki/CZ_1_13_2/phase3/maintenance php runJobs.php 2>&1 > ~/runJobs.log1& php runJobs.php 2>&1 > ~/runJobs.log2& php runJobs.php 2>&1 > ~/runJobs.log3& php runJobs.php 2>&1 > ~/runJobs.log4& php runJobs.php 2>&1 > ~/runJobs.log5& php runJobs.php 2>&1 > ~/runJobs.log6& php runJobs.php 2>&1 > ~/runJobs.log7& php runJobs.php 2>&1 > ~/runJobs.log8& php runJobs.php 2>&1 > ~/runJobs.log9& php runJobs.php 2>&1 > ~/runJobs.log10& php runJobs.php 2>&1 > ~/runJobs.log11& php runJobs.php 2>&1 > ~/runJobs.log12& php runJobs.php 2>&1 > ~/runJobs.log13& php runJobs.php 2>&1 > ~/runJobs.log14& php runJobs.php 2>&1 > ~/runJobs.log15& php runJobs.php 2>&1 > ~/runJobs.log16& php runJobs.php 2>&1 > ~/runJobs.log17& php runJobs.php 2>&1 > ~/runJobs.log18& php runJobs.php 2>&1 > ~/runJobs.log19& php runJobs.php 2>&1 > ~/runJobs.log20& wait
This script creates 20 log files in the user's home page that document the jobs activity.
After running runJobs, the statistics page will be out of date. To update it, execute:
cd /usr/local/src/mediawiki/CZ_1_13_2/phase3/maintenance php initStats.php
Loading the clone database from a postgres database dump file
Loading the clone database from a CZ daily dump file is extremely slow. On a dual 1.8 GHz processor with 4 GB of memory, loading one of the CZ daily dump files from December 2009 took 80 hours (that is no misprint: ~3.25 days!). Fortunately, there is a way to work around this problem.
Once the daily data dump file is loaded into the CZ clone, it is possible to dump the postgres cz database and make it available. Using the postgres database dump reduces the time to load the CZ clone database from 80 hours to a little over 21 minutes (for the postgres database of 01-18-2010 on a dual 1.8 GHz processor with 4 GB of memory).
First, we need to download the most recent public postgres database dump. Click on the following link — postgres database dumps. Select the most recent dump file. Files are named cz_db_dump_<date>.gz, where <date> is the date of the dump. For example, the dump of Jan 18, 2010 is named cz_db_dump_1_18_2010.gz. Click on the file you wish to download and a dialog window should appear asking you where to save it. Save it to your home directory (or save it anywhere and copy/move it to your home directory). It may take some time to download the file, since the database dumps are fairly large. For example, the file cz_db_dump_1_18_2010.gz is 174 MB.
To use the postgres database dump, it is necessary to first re-create the cz database. This is simple. First, delete the current cz database using pgAdmin III. From the main window, place the mouse cursor over the cz icon (you may need to double click the localhost icon to reveal the cz icon). Then right click and select _Delete/Drop (see figure). A window pops up asking you whether you are sure you wish to drop the cz database. Click OK. Now recreate the cz database. In pgAdmin III right click Databases and select New Database. In the name field of the window that appears enter cz and select cz from the drop down menu for the owner. Click OK.
Having re-created the cz database, load the database dump by executing the following command in a terminal window:
gunzip -c cz_db_dump_<date>.gz | psql cz
Of course, you must substitute the date of the dump file for <date>. For example, to load the dump file cz_db_dump_1_18_2010.gz execute:
gunzip -c cz_db_dump_1_18_2010.gz | psql cz
Since you deleted the cz database you created originally, the user name and password for the first user was also deleted. The user name and password for the first user in the postgres dump wiki is "Wikiadmin" and "Wikiadminpw". This user has bureaucrat privileges. You can use the preferences tab to change the password if you wish.
Keeping up to date with CZ wiki changes
Whether you loaded your CZ clone data using importDump.php or using the postgres database dump, the data on your clone will not track the changes made to the Citizendium wiki unless you periodically update it. This is achieved using the CZ daily dump file. You may update your clone's data as frequently or as infrequently as you see fit. In fact, if you set up your clone only as the foundation for a local development environment, you may decide to never update its data. On the other hand, if you set up your clone in order to work on articles when a connection to the internet is unavailable, then you may wish to update your clone's data frequently.
Updating the clone's data uses importDump.php. This utility is designed so it changes only those pages that are out of date. So, if you use it often to update your clone, it's execution time should be reasonable. For example, running on a dual 1.8 GHz processor with 4 MB of memory, importing the daily dump 4 days subsequent to the date of the postgres database dump took about 30 minutes. On the other hand, importing the data from a daily dump made 6 weeks after an initial import took about 10.5 hours. It is up to you to decide the tradeoffs between update frequency and execution time.
To update your clone's data use the instructions given in Loading the clone database from a CZ daily dump file.