Installation Manual for 2.15 on Ubuntu 20.04 Server

These instructions cover the installation of the Ubuntu Server 20.04 LTS operating system and WeBWorK 2.15 from scratch.

You can also look at Installation_Manual_for_2.15_on_Ubuntu_20.04_Desktop for instructions on the installation of Ubuntu Desktop 20.04 LTS operating system and WeBWorK 2.15. The main difference between the Desktop and Server versions is that the Desktop version includes a GUI and the Server version does not.

Other installation methods (which are faster and easier) include using the WeBWorK Virtual Machine Image (see Installing WeBWorK from a Virtual Machine Image) which has been tested on VMware, VirtualBox and QEMU/KVM and using the WeBWorK AMI if you want to run your server on AWS (see WeBWorK Amazon Machine Images (AMI's)).

Notation

First some short comments on notation we will be using. We will use <key> to indicate that you should press a specific key (e.g. <Enter>, <Tab>, <F12>, etc.). Sometimes we will also use e.g. <wwadmin password> to indicate you have to enter the wwadmin password.

^ will indicate the <Ctrl> key so e.g. ^X is really shorthand for <Ctrl> <X>, i.e. press the Ctrl key and hit the X key.

Installing the Ubuntu 20.04 Server Operating System

Installation DVD

Obtain the Server Edition installation DVD/CD set. Connect to http://www.ubuntu.com/ for information. For example you can download an ISO image of the installation DVD and then burn your own installation DVD. You want the file ubuntu-20.04-live-server-amd64.iso. Note 1: We recommend you use the "Long Term Support" (LTS) version of Ubuntu which is currently version 20.04; the next LTS release will be version 22.04 to be released in April, 2022. Note 2: We recommend you actually use the latest 20.04 iso file which is currently ubuntu-20.04-live-server-amd64.iso. Choose whichever download site works best for you; I have had good luck with http://mirrors.kernel.org/ubuntu-releases/ If you download the ISO image, make sure that you verify the integrity of the downloaded file by comparing the MD5 checksum of the downloaded file with the MD5 checksum listed at https://help.ubuntu.com/community/UbuntuHashes or at the download site (e.g. http://mirrors.kernel.org/ubuntu-releases/20.04). These instructions will assume you have the ubuntu-20.04-live-server-amd64.iso installation DVD but installing from an alternate DVD, a commercial DVD/CD set or from the net should be essentially identical.

You will want to have you computer connected by ethernet to the internet for the installation. Place the installation CD in your DVD/CD drive and reboot your computer from the DVD drive. You may have to press a key (e.g. <F12>) during the boot process to bring up a boot menu which will allow you to select booting from the DVD. Or you many have to edit the BIOS to select the DVD as the first boot device. See Step 3 in https://linuxconfig.org/ubuntu-20-04-server-installation for help. Or maybe you are installing the iso image into a virtual machine.

After the system boots you will be presented with a series of 8 steps or panels. In the panels, generally you will use <Tab> or the up and down arrows to navigate, the <Space> bar to check a box and the <Enter> key to confirm an action.

Step 1
On the first install panel keep English as the selected language and press <Enter>.
Step 2
press <Enter> to accept the default keyboard layout
Step 3
press <Enter> to accept the default mirror for Ubuntu

Optional Configurations

If you will have a large number of users (say over a 1,000) and/or a slow server, you may want to consider the first two optimizations. They are independent but related and deal with how WeBWorK handles various temporary and static files. We call these two options Optional A and Optional B. The third option, Optional C, gives greater security and should be used for any server students are using. Optional D, which installs and configures the R server for statistical computing, should be implemented if you plan to assign probability or statistics problems. Finally Optional E sets up log rotation for WeBWorK's timing log which can grow huge if you have a large number of users.

Optional A creates a separate partition (or directory) on which are stored all of WeBWorK's "temporary" files. These are mostly small files such as png images of equations, pdf files, etc. that may be reused but if they are not present (e.g. if they get deleted) they will be seamlessly regenerated on the fly. There is no reason to backup such files and having them in a separate partition or directory means that it is easier and faster to backup other partitions and skip backing up unnecessary files. Even if you do not want to set up a separate partition for this, it is very convenient to at least set up a separate directory for these temporary files and we recommend that you do so. But this is something you do after WeBWorK is installed.

Optional B installs and configures a lightweight webserver. Apache is a very standard and powerful webserver which we use to serve WeBWorK pages. However its child processes use a lot of resources (e.g. memory). When serving static files and images, a much lighter weight webserver can be used. This can substantially reduce the load on a heavily used server.

Optional C configures Apache (and optionally lighttpd if installed) so that access to WeBWorK will be through an encrypted connection (SSL) with an https: URL.

Optional D installs and configures the R server for statistical computing. Many statistics and probability problems (mostly contributed by faculty at the University of British Columbia) require this.

Optional E sets up log rotation for WeBWorK's timing log.

Except for creating a separate partition, we will wait until WeBWorK is installed and tested before implementing these options. We mention them here because the next step is partitioning the disks.

Partition disks

Step 4
Next comes the Guided storage configuration page. You should be able to accept the defaults unless you want to follow Optional A and/or create separate partitions for various directories. There is a lot of information on the web if you don't want to accept the default partition set up. If you want to implement Optional A follow the directions below. Note that if you only want to implement Optional A with a separate directory, not a separate partition, you don't have to do anything special with partitions at this point. Just select Done and press <Enter>.

Optional A: The default partitioning scheme creates just two partitions, a root (/) partition and a swap partition. Actually as of Ubuntu version 18.04, the default is to create a swap file instead of a swap partition. Here we will create those and an additional partition for WeBWorK's temporary files.

On the Guided storage configuration page select Custom storage layout (by using the <Space>) bar, then select Done and press <Enter>. This will bring you to the Storage configuration page where you can choose a new partition table, resize partitions, etc.

Now you have to decide how to allocate your disk space. For WeBWorK's temporary files 25 GB for every 1,000 students should be ample. For recommendations on the size of a swap partition and for help setting up partitions, etc. consult the web.

Continue Initial Installation

Step 5
Confirm the partition you want to install Ubuntu in, select Done and press <Enter>. Now confirm the changes ("Confirm destructive action") by selecting

Continue and pressing <Enter>.

Step 6
The next panel is "Profile setup". What you fill in here is a privileged user so you might want to enter a special administrative account rather than your normal account. We suggest <wwadmin> but you can use whatever you want.
• Enter the information for "Your name"
• Enter the information for "Your Computer's name"
• Enter the information for "Pick a username" In these instructions, we will call this name <wwadmin>.
• Enter the information twice for the password. In these instructions, we will call this password the <wwadmin password>. Do not forget what you enter here.
• Then select Done and press <Enter>.

Step 7
Select "Install OpenSSH server". Then select Done and press <Enter>.
Step 8
Do not select any server snaps. Then select Done and press <Enter>.

Now sit back and relax while the installation takes place. This should be pretty quick. When you finally see "Reboot", press <Enter>.

Continue Installation

The system will ask you remove the DVD. Just press <Enter>. Note that when installing on VirtualBox, it froze at:

[ OK ] Reached target Cloud-init target


If this happens to you, go into "Machine" at the top in VirtualBox and press "ACPI Shutdown" and then just start it again.

Terminal Window Notation and Use

Before installing and configuring additional software, we need to talk about the terminal window.

In the terminal window some commands will have to be run as root whereas others should be run as a regular user. We will use # to indicate that the command is to be run as root e.g.

# perl -MCPAN -e shell


and $to indicate that the command is to be run as a normal user e.g. $ cp .bashrc .bashrc.bak1


To execute the above commands you have to hit <Enter>. We'll just assume this. After executing a command, often the system will respond with text (sometimes a lot of text!) which we will usually not repeat below. We only give the commands that you should execute.

The bash shell which you will be using has a number of very convenient features.

One is command and file name completion. If you are typing (e.g. ch) and hit <tab> bash will complete the command or filename if it is unambiguous (or more precisely it will complete as much as possible). If there are multiple possibilities (as in the case of ch) nothing will happen (except you may hear a beep) and you can type more letter(s) and hit <tab> again. Or you can hit <tab> a second time and you will see a list of all possible completions. E.g. entering ch<tab><tab> gives a list of possible completions and ch<tab>gr<tab> (or chgr<tab>) gives chgrp, the change group command. This is very fast and convenient and it also leads to fewer typing errors.

Another useful shortcut is the command history. Using the up and down arrow keys will bring up previous commands which can be edited and then executed. If you are repeating a command or entering a command which is similar to a previous one, this is very useful.

For our first terminal window task create a downloads directory where we will keep copies of downloaded software.

$cd$ mkdir downloads


Running commands as root

By default Ubuntu has no password set for the root user. To gain root access you have to use the sudo command with the <wwadmin password>. This is the password you set for the first user which we called the WeBWorK administrator <wwadmin> while installing Ubuntu.

To run commands as root use the sudo command

$sudo <command> [sudo] password for wwadmin: <wwadmin password>  After you enter the password the command is executed. For a certain period (maybe 5 minutes) you can execute additional sudo commands without reentering <wwadmin password>. A log of all sudo commands is kept in /var/log/auth.log . You can also use sudo to become root and get the root prompt #. To do this run $ sudo su
#


When you want to exit the root prompt and return to being a regular user, do the following

# exit
exit
$ Shutting down the system To shut down the system, use the shutdown command. sudo shutdown --help [sudo] password for wwadmin: <wwadmin password>  will show you all options. For example shutdown -h 1 will shut the system down in 1 minute. All these commands must be run as root. E.g. sudo shutdown -h 1 [sudo] password for wwadmin: <wwadmin password>  Accessing Your Server Remotely At this point you can login your server from a remote location using SSH (non secure telnet and FTP are not allowed but secure SSH and SFTP are). You can do all of the remaining installation from a remote location if you wish. The advantage of doing this is that you can copy commands from these instructions (with copy from the Edit menu or ^C) and paste them into a terminal window (with paste from the Edit menu list or <Shift> <Ctrl> <V> or <Shift> <Insert> depending on your application). This is an excellent way to use these instructions since it is fast and insures commands are entered correctly (just be careful to read before you run the command and replace things like database_password with the correct code in the few places such things occur). Assuming your network has been set up automatically, type $ ip addr show


and you will see something like

...
inet 192.168.76.128/24 brd 192.168.76.255 scope global dynamic ens33
...


Here 192.168.76.128 is the ip address of your server. Note that you should ignore the loopback address 127.0.0.1.

If your network has not been set up automatically, you will have to set it up manually. Doing that is beyond the scope of these instructions.

I'm assuming you are logging in remotely so that you can easily copy and paste commands from these instructions into your terminal application. After login, you may see a message about update that can be installed. To do this, run the command

sudo apt update


to see the list of available updates and then to install them

sudo apt upgrade
...
Do you want to continue? [Y/n]


and just press <Enter> to accept the default "Y".

Set the Timezone for your server

To find out what timezone your server is set to run the command

$timedatectl  and you will probably see ... Time zone: Etc/UTC (UTC, +0000) ...  which is unlikely to be where you live. The timezone naming convention uses a “Region/City” format and to find the correct one for your location run the command $  timedatectl list-timezones


Look through the list and find your timezone, e.g. "America/New_York". Then set the timezone (you have to be root), e.g.

$sudo timedatectl set-timezone America/New_York [sudo] password for wwadmin: <wwadmin password>  and then $ timedatectl


to check it was set correctly.

Ubuntu Software Packages

First let's make sure our packages sources are up tp date. Run

$sudo apt update [sudo] password for wwadmin: <wwadmin password>  Our next task is to install a number of Ubuntu software packages. Here is the list of Ubuntu packages that need to be installed. 1. apache2 2. curl 3. dvipng 4. gcc 5. git 6. libapache2-request-perl 7. libarray-utils-perl 8. libcrypt-ssleay-perl 9. libdatetime-perl 10. libdancer-perl 11. libdancer-plugin-database-perl 12. libdata-dump-perl 13. libdbd-mysql-perl 14. libemail-address-xs-perl 15. libemail-sender-perl 16. libexception-class-perl 17. libextutils-xsbuilder-perl 18. libfile-find-rule-perl-perl 19. libgd-perl 20. libhtml-scrubber-perl 21. libiterator-perl 22. libiterator-util-perl 23. libjson-perl 24. liblocale-maketext-lexicon-perl 25. libmime-tools-perl 26. libmoox-options-perl 27. libnet-ip-perl 28. libnet-ldap-perl 29. libnet-oauth-perl 30. libossp-uuid-perl 31. libpadwalker-perl 32. libpath-class-perl 33. libphp-serialization-perl 34. libpod-wsdl-perl 35. libsoap-lite-perl 36. libsql-abstract-perl 37. libstring-shellquote-perl 38. libtemplate-perl 39. libtext-csv-perl 40. libtimedate-perl 41. libuuid-tiny-perl 42. libxml-parser-easytree-perl 43. libxml-parser-perl 44. libxml-writer-perl 45. make 46. mysql-server 47. netpbm 48. preview-latex-style 49. texlive 50. texlive-latex-extra 51. unzip To install all of these packages in one fell swoop, first become root: $ sudo su


and then run the command (obviously you want to use cut and paste)

# apt install apache2 curl dvipng gcc git libapache2-request-perl \
libarray-utils-perl libcrypt-ssleay-perl libdancer-perl libdancer-plugin-database-perl \
libdata-dump-perl libdatetime-perl libdbd-mysql-perl libemail-address-xs-perl libemail-sender-perl \
libexception-class-perl libextutils-xsbuilder-perl libfile-find-rule-perl-perl \
libgd-perl libhtml-scrubber-perl libiterator-perl libiterator-util-perl libjson-perl \
liblocale-maketext-lexicon-perl libmime-tools-perl libmoox-options-perl libnet-ip-perl libnet-ldap-perl \
libphp-serialization-perl libpod-wsdl-perl \
libsoap-lite-perl libsql-abstract-perl libstring-shellquote-perl libtemplate-perl \
libtext-csv-perl libtimedate-perl libuuid-tiny-perl libxml-parser-easytree-perl \
libxml-parser-perl libxml-writer-perl make mysql-server netpbm \
preview-latex-style texlive texlive-latex-extra unzip



If prompted, you can always accept the default (hit <Enter>). When the process finishes, enter

# exit


Installing Perl Modules

We now have to install several additional Perl modules which unfortunately are not available from the Debian package system.

Testing Perl Modules

To test if a Perl module is installed and working on your system, issue the following command, replacing Module with the name of the module:

$perl -MModule -e 'print "installed!\n"'  If the module is installed you will see installed!. If not you will see at lot of gibberish. E.g. at this stage in our installation process CPAN is installed and Statistics::R::IO is not so $ perl -MCPAN -e 'print "installed!\n"'


yields

installed!


and

$perl -MStatistics::R::IO -e 'print "installed!\n"'  yields Can't locate Statistics/R/IO.pm in @INC ...  You can check the version of an installed module by the following command, replacing Module with the name of the module:  perl -MModule -e 'print "$Module::VERSION\n"'


For example for the GD.pm module

 perl -MGD -e 'print "$GD::VERSION\n"'  Installing Additional Perl Modules from CPAN First we will set up CPAN. For this you have to be root. $ sudo su
# perl -MCPAN -e shell


Since this is the first time you are using CPAN it will ask you Would you like me to configure as much as possible automatically? Answer Yes (the default) and that should be it. If you are asked for a method to use, choose sudo. If you are asked choose some CPAN mirror sites, you can just answer Yes (the default).

Now install the following modules

cpan> install Statistics::R::IO


and in case you are prompted accept all defaults by just hitting <Enter>. Note that with more than one module to install, we just list them after install separated by spaces.

When you again see the

cpan>


prompt enter

cpan> exit
#
exit
$ Upgrading CPAN This isn't really necessary, but you will see messages about upgrading CPAN. If you want to upgrade, do the following $ sudo su
# perl -MCPAN -e shell


Now we update CPAN itself

cpan> install Bundle::CPAN


and always hit <Enter> to accept the defaults when prompted (e.g. when you see "Enter arithmetic or Perl expression: exit"). This can be a long process, please be patient. When you again see the

cpan>


prompt enter

cpan> reload cpan
cpan> o conf commit


and then quit

cpan> exit
#
exit
$ Editing Files You can use your favorite editor (e.g. vim) but we will give instructions assuming you are using the simple text editor nano. Apache 2 and mod_perl We enable the info module. Acting as root in a terminal window enter $ sudo su
# a2enmod info


Next we make a copy of the configuration files we will be editing for safekeeping.

# cd /etc/apache2/mods-available

# cp info.conf info.conf.bak1
# cp status.conf status.conf.bak1

# exit
$ info.conf and status.conf Now we will edit configuration files info.conf and status.conf to allow us to view information about the setup and performance of the web server. Note that this is not absolutely necessary but it can be very useful. You can use your favorite editor but we will give instructions assuming you are using nano. Note that you have to be root to edit these files. First we edit info.conf $ cd /etc/apache2/mods-available
$sudo nano info.conf [sudo] password for wwadmin: <wwadmin password>  I suggest you allow access to server information from e.g. your department domain. To do this uncomment (i.e. remove the # from)  Require ip 192.0.2.0/24  and then replace 192.0.2.0/24 by .math.yourschool.edu where of course you should edit .math.yourschool.edu appropriately. Then save the file and quit (press ^X, Y, <Enter>). Now we edit status.conf $ cd /etc/apache2/mods-available
$sudo nano status.conf [sudo] password for wwadmin: <wwadmin password>  Edit the  #Require ip 192.0.2.0/24  line just as you did for info.conf. Then save the file and quit. And restart Apache so that the above changes take effect $ sudo systemctl restart apache2


Fully Qualified Domain Name

Now we have to set your server's fully qualified domain name. Note that if your network was set up automatically via DHCP, your server's fully qualified domain name should already be set up. You can check by running the hostname commands below.

Run the command

sudo hostnamectl set-hostname webwork


where of course you should replace webwork by whatever your server's name is.

You can check these settings by running the commands

$hostname --fqdn  and $ hostname


The first gives the server's fully qualified domain name (e.g. webwork.mydepartment.myschool.edu) and the second the server's name (e.g. webwork).

Note that if your server can not find its fully qualified domain name, certain tools (such as the Synaptic Package Manager) may not start.

Test Apache

Now restart Apache

$sudo systemctl restart apache2 password:<wwadmin password>  or $ sudo apache2ctl restart


or

$sudo service apache2 restart [sudo] password for wwadmin: <wwadmin password>  and test your server by connecting to your server ("http://yourserver.yourschool.edu") from a browser on a remote machine. You should see the Apache2 Ubuntu Default Page indicating that Apache is running. You can check Apache's status by connecting to "http://yourserver.yourschool.edu/server-status" from a browser on a remote machine in the math.yourschool.edu domain. Further test Apache by connecting to "http://yourserver.yourschool.edu/server-info" from a browser on a remote machine in the math.yourschool.edu domain and you will see a page listing various information about Apache. In particular under Server Settings you should see Server Version: Apache/2.4.41 (Ubuntu) mod_apreq2-20090110/2.8.0 mod_perl/2.0.11 Perl/v5.30.0  (or something very similar) indicating that both mod_apreq2 and mod_perl are installed. If you have problems now or in the future, a good first thing to do is to look at the Apache error log which is located at /var/log/apache2/error.log. In the directory /var/log/apache2/ you can "less" through the error log (less error.log), look at the last few entries (tail error.log) or run the command tail -f error.log which will display new error messages as they are appended to the file. Use ^C to break out of tail -f . For example if you didn't set up access to "server-info", you can run tail /var/log/apache2/error.log  and in the output you should see something similar to ...[mpm_event:notice] [pid 48932:tid 140494759132224] AH00489: Apache/2.4.41 (Ubuntu) mod_apreq2-20090110/2.8.0 mod_perl/2.0.11 Perl/v5.30.0 configured -- resuming normal operations  indicating that both mod_apreq2 and mod_perl are installed and running. Checking MySQL First check that MySQL is running by $ sudo mysql


You should see something very similar to

Welcome to the MySQL monitor.  Commands end with ; or \g.
Your MySQL connection id is 8
Server version: 8.0.20-0ubuntu0.20.04.1 (Ubuntu)
...

Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.


Enter exit to exit

mysql> exit
Bye
$ Checking Apache If you have not done this yet connect to your server (e.g. " "http://webwork.math.yourschool.edu/" or use it's ip address) from a browser on a remote machine. You should see the Apache2 Ubuntu Default Page indicating that Apache is running. MySQL Security and Performance Issues As initially set up, MySQL is an open system. There are anonymous accounts with full privileges for some databases and other issues. We will fix that now. Run the command sudo mysql_secure_installation [sudo] password for wwadmin: <wwadmin password>  You can setup the VALIDATE PASSWORD plugin if you wish. I chose not to. Next you will be asked to enter a New password for the MySQL "root" user. Enter your chosen MySQL root password. You have to enter a password but it will not be used. In Ubuntu systems running MySQL 5.7 (and later versions), the root MySQL user is set to authenticate using the auth_socket plugin by default rather than with a password. You almost certainly want to answer 'Y' to all questions except possibly for the first one which asks if you want to set up a VALIDATE PASSWORD plugin. Now test that all is well: $ sudo mysql


You should see

Welcome to the MySQL monitor ...
mysql>


Now lets check the MySQL users. There are five accounts. To see the accounts, do the following

mysql> SELECT user,authentication_string,plugin,host FROM mysql.user;


You will see a table with five users (debian-sys-maint, mysql.infoschema, mysql.session, mysql.sys and root). You should see that one of the five users has a valid password (which will be displayed in encrypted form) and root is authenticated by a socket. You can ignore the mysql.session, mysql.sys and mysql.infoschema users.

Now exit MySQL

mysql> exit
Bye
$ This last step is optional since currently most of WeBWorK's MySQL tables are now creating using the MyISAM engine (by default -- this can be changed in the config files) rather than the MySQL default engine. Now we change MySQL's default engine. The default engine is InnoDB as of MySQL 5.5.5 (MyISAM before 5.5.5) but (at least on some hardware) InnoDB seems to be 50-100 times slower than MyISAM. So we will change the default MySQL engine from InnoDB to MyISAM. Note that this change only applies to new tables, tables already constructed will continue to use InnoDB. But we haven't created any WeBWorK tables so we don't have to change the engine for any existing tables. Note: There is a report of a problem using MyISAM by a French speaking professor who changed MySQL's default character set to utf8 (see http://webwork.maa.org/moodle/mod/forum/discuss.php?d=3174). If you change MySQL's default character set, you should research what is the best engine to use. To change the default MySQL engine from InnoDB to MyISAM do the following: $ cd /etc/mysql/mysql.conf.d
$sudo cp mysqld.cnf mysqld.cnf.bak1 [sudo] password for wwadmin: <wwadmin password>$ sudo nano mysqld.cnf


Search for [mysqld] and under "Basic Settings" above the "user = mysql" line add the line

default-storage-engine = myisam


Then save the file and quit. Restart Mysql

$sudo /etc/init.d/mysql restart [sudo] password for wwadmin: <wwadmin password>  or $ sudo service mysql restart


$sudo mysql [sudo] password for wwadmin: <wwadmin password>  You will see Welcome to the MySQL monitor ... mysql>  If you show MySQL engines, you should see that MyISAM is listed as the default engine. mysql> SHOW ENGINES\G  Now exit mysql> exit Bye$


We are finally at the point where we can start downloading and installing WeBWorK. We will use Git to download WeBWorK from Github. This is easy and it will also make it easy to update the system in the future. Note that the following are rather long commands; it is much easier to copy (^C) them from this document and paste (<Shift> <Ctrl> <V> or <Shift> <Insert>) them in a terminal window

$cd$ cd downloads
$git clone git://github.com/openwebwork/webwork2.git$ git clone git://github.com/openwebwork/pg.git
$git clone git://github.com/openwebwork/webwork-open-problem-library.git$ git clone git://github.com/mathjax/MathJax.git


Important Note. The above commands retrieve the master branch which gives the latest stable release of the software package (webwork2, pg, etc.) with bug fixes. If a stable release newer than 2.15 exists, that will be downloaded and these instructions may be a little out of date. So it is a good idea to check before downloading. The best way to do that is to look at https://github.com/openwebwork/webwork2/blob/master/VERSION and https://github.com/openwebwork/pg/blob/master/VERSION.

The first and second download gives you the latest released versions. The third download contains the WeBWorK Open Problem Library (OPL) which is the new name for the original National Problem Library (NPL). The NPL has been renamed to reflect its growing international content. Your system will be loaded with many thousands of WeBWorK problems (over 30,000 currently). The fourth download is MathJax which is one of several options WeBWorK has to display mathematics online.

The main information page about the OPL is available at http://webwork.maa.org/wiki/Open_Problem_Library

The main information page about MathJax is available at http://www.mathjax.org/

Installing WeBWorK

Note the instructions below assume you are installing WeBWorK from scratch. If you are just upgrading WeBWorK, especially if you already have existing WeBWorK courses, see Upgrading WeBWorK.

Move the System into the Required Directories

As root create a webwork directory under /opt, change the ownership of the webwork directory to wwwadmin and move directories there.

$cd$ cd downloads
$sudo mkdir /opt/webwork <wwadmin password>$ sudo chown wwadmin:wwadmin /opt/webwork
$mv webwork2 /opt/webwork/$ mv pg /opt/webwork/


Move the mathjax directory to its proper location

$mv MathJax /opt/webwork/  Now create the courses and libraries directories under webwork and copy and move content there $ mkdir /opt/webwork/courses
$mkdir /opt/webwork/libraries$ mv webwork-open-problem-library /opt/webwork/libraries/
$cd /opt/webwork/webwork2/courses.dist$ cp *.lst /opt/webwork/courses/
$rsync -a modelCourse /opt/webwork/courses/  Setting Permissions The PG installation directory and files should be owned by wwadmin and not writable by other users: $ cd /opt/webwork/pg
$chmod -R u+rwX,go+rX .  Most WeBWorK directories and files should also be owned by wwadmin and not writable by other users: $ cd /opt/webwork/webwork2
$chmod -R u+rwX,go+rX .  Certain data directories need to be writable by the web server. These are DATA, courses, htdocs/tmp, logs, and tmp. It is convenient to give WeBWorK administrators access to these directories as well, so they can perform administrative tasks such as removing temporary files, creating and editing courses from the command line, managing logs, and so on. We will create a new group called wwdata, containing both the WeBWorK administrators and the web server. Run the commands $ sudo addgroup wwdata
$sudo adduser wwadmin wwdata  If there are other users who will also be administering WeBWorK files, now is a good time to add them. To do this use the sudo adduser 'username' command. And remember to add the new account to the wwdata group as above. Now add the Apache2 webserver (which runs as www-data) to the wwdata group: $ sudo adduser www-data wwdata


You can check that this succeeded in a terminal window by entering

$id wwadmin  and then you should see wwdata listed under groups. Also $ id www-data


should show wwdata listed under groups.

Now we make the WeBWorK directories that need to be writable by the web server have wwdata as their group. The following are rather long commands; you might want to copy them and paste them into your terminal window rather than typing them.

$sudo su [sudo] password for wwadmin: <wwadmin password> # cd /opt/webwork/webwork2/ # chgrp -R wwdata DATA ../courses htdocs/tmp htdocs/applets logs tmp /opt/webwork/pg/lib/chromatic # chmod -R g+w DATA ../courses htdocs/tmp htdocs/applets logs tmp /opt/webwork/pg/lib/chromatic # find DATA/ ../courses/ htdocs/tmp logs/ tmp/ -type d -a -exec chmod g+s {} \; # exit$


Compile color.c

$cd /opt/webwork/pg/lib/chromatic$ gcc color.c -o color


You may see some warning messages which you can safely ignore.

Configuring the Shell

To make working with WeBWorK easier, there are a couple of changes you can make to your shell environment.

Add the WeBWorK bin directory to your path. This will allow you to run WeBWorK command-line utilities without typing the full path to the utility. Goto your home directory and backup your .bashrc file

$cd$ cp .bashrc .bashrc.bak1


Now edit .bashrc

$nano .bashrc  After the last line add the three lines: export PATH=$PATH:/opt/webwork/webwork2/bin
export WEBWORK_ROOT=/opt/webwork/webwork2
export PG_ROOT=/opt/webwork/pg


Then save the file and Quit.

Close your Terminal Window and open a new one so the above changes take effect. You can check that they have by

$echo$PATH
$echo$WEBWORK_ROOT
$echo$PG_ROOT


Checking Module Dependencies

WeBWorK includes a script called check_modules.pl (in the directory /opt/webwork/webwork2/bin) that verifies that the needed programs and Perl modules are installed on your system. Run this script to make sure you have installed the required programs and Perl modules.

$check_modules.pl apache2  Scroll up and look through the listing. It should find everything. If something is missing (flagged by **), look back through these instructions to find where it should have been installed and install it. You may see some warning messages like Prototype mismatch: sub main::from_json: none vs ($@) at (eval 188) line 2.
Prototype mismatch: sub main::to_json: none vs ($@) at (eval 188) line 2.  This is a known bug in libjson-perl and can be safely ignored. Now we check that all necessary LaTeX packages have been installed. Run the commands $ cd
$pdflatex /opt/webwork/webwork2/bin/check_latex.tex  and look for missing packages (you can ignore "No file check_latex.aux."). Configuring WeBWorK Making Copies of the Distribution Configuration Files Before configuring the system, you must make local copies of the site.conf and localOverrides.conf configuration files, located in /opt/webwork/webwork2/conf/ . $ cd /opt/webwork/webwork2/conf
$cp site.conf.dist site.conf$ cp localOverrides.conf.dist localOverrides.conf


System Configuration

Most WeBWorK configuration is done in the files /opt/webwork/webwork2/conf/site.conf and /opt/webwork/webwork2/conf/localOverrides.conf. These files provide system-wide configuration settings, and defaults for course settings. Any setting in these files can be overridden in the course.conf file for a particular course. To override a setting for a course, just put the new setting (using the same syntax as is in localOverrides.conf) in the course.conf file. An instructor can only edit the course.conf file herself (for her own course) if she has "admin" privilege which by default instructors do not have. But most things instructors may want to customize and many others (language, timezone, permissions, display modes, email, ...) can be set using the Course Configuration page from within the course and such setting override those in the configuration files.

Actually there are three main configuration files, site.conf, defaults.config and localOverrides.conf. The reason there are three configuration files is to make upgrading WeBWorK easier.

• site.conf: This file contains global variables which are required for basic configuration. It will not be overridden when you update WeBWorK but its distribution version, site.conf.dist will be.
• defaults.config: This file contains initial settings for many customizable options in WeBWorK. Do not edit defaults.config. It will be overridden next time you upgrade.
• localOverrides.conf This is where you should add all local customizations. It will not be overridden when you update WeBWorK but its distribution version, localOverrides.conf.dist will be.

There are several options that must be set for WeBWorK to work with your system. The rest of the file consists of customization options.

Edit the site.conf file

Now edit site.conf

$cd /opt/webwork/webwork2/conf$ nano site.conf


First we have to add information about the Apache2 server setup. Search for server_root_url and edit the line so that it reads:

$server_root_url = 'http://yourserver.yourschool.edu';  where of course you should edit yourserver.yourschool.edu appropriately. The "http://" is important. If you are running a secure server (i.e., using SSL - really TLS), the url should start with https: but you should wait to make that change until after WeBWorK is up and running and you set up SSL. In that case the "https://" is important. We need to set a password that WeBWorK uses when it communicates with the MySQL database. Note that this is not the same as the <mysql root password> which is the unused password we set when securing MySQL. Search for database_password and replace the line $database_password = "passwordRW";


by

$database_password = 'database_password';  where of course you should replace "database_password" with your own password but do use the single quotes. Remember this password as we will need it below. WeBWorK sends mail in three instances. The PG system sends mail to report answers to questionnaires and free-response problems. The mail merge module is used to send mail to course participants, i.e. to report scores. The feedback module allows participants to send mail to course instructors. To send mail, WeBWorK needs the address of an SMTP server. Normally you will use the address of your school's SMTP server. If the local machine is running an SMTP server, use localhost. IMPORTANT: Our instructions above did not install an SMTP server so you will have to install and configure one if you do not use your school's SMTP server. When connecting to the SMTP server, WeBWorK must also send an email address representing the sender of the email (this has nothing to do with the From address on the mail message). Edit the lines $mail{smtpServer} = ;  # e.g. 'mail.yourschool.edu' or 'localhost'
$mail{smtpSender} = ; # e.g. 'webwork@yourserver.yourschool.edu'  entering the appropriate information. Be sure to use single quotes and NOT double quotes around email addresses otherwise Perl will treat @ as an array variable. WeBWorK uses the DateTime module. DateTime is supposed to be able to determine the local timezone itself without you having to enter it but this often fails so it is best to just set it here. For is a list of timezones recognized by DateTime run the command timedatectl list-timezones  These timezones are more refined than standard time zone usage in that they include switches to daylight savings time (e.g. some parts of a time zone may make the switch and others may not). For example if your server is in the eastern US, on the list you will see America/New_York and you should enter $siteDefaults{timezone} = "America/New_York"; which is the default. Read the documentation in this section of the the site.conf file for more information on selecting time zones and formatting dates.

Search for $siteDefaults{timezone} and enter your local timezone if it is not correct. Then save the file and Quit. The defaults.config file If you want WeBWorK questionnaires or similar things from different courses to be mailed to a central person or persons (e.g. the WeBWorK administrator), in defaults.config, you will see the lines $mail{allowedRecipients}     = [
#'prof1@yourserver.yourdomain.edu',
#'prof2@yourserver.yourdomain.edu',
];


But we are not supposed to edit the defaults.config file, so if we want to do this, we will copy this to localOverrides.conf and edit it appropriately. Note that we should move this setting to the site.conf file.

Edit the localOverrides.conf file

$cd /opt/webwork/webwork2/conf$ nano localOverrides.conf



As we said above, if you want WeBWorK questionnaires or similar things from different courses to be mailed to a central person or persons (e.g. the WeBWorK administrator), in localoverrides.config, add and then edit the lines

$mail{allowedRecipients} = [ #'prof1@yourserver.yourdomain.edu', #'prof2@yourserver.yourdomain.edu', ];  removing the # and using the professor(s) actual email address(es). In order to have professors from individual courses receive such email, this should be set in course.conf (which you find in the course directory) to the addresses of professors of each course. Note that the settings in course.conf override the settings in site.conf, default.conf and localOverrides.conf so if in addition you want e.g. the WeBWorK administrators to receive copies, you have to add them as well. Then save the file and Quit. After you have used WeBWorK for awhile, you may want to change the default header files in defaults.config. Search for $webworkFiles{screenSnippets}{setHeader} and $webworkFiles{hardcopySnippets}{setHeader} in localOverrides.config. Settings in the conf and config files affect all WeBWorK courses. You can override any setting in these conf and config files for an individual course by putting the local setting in course.conf. Also now there are several versions of the classlist editor, homeworkset editor, library browser and pgproblem editor. Which ones are available in a course is determined by the settings for %showeditors in localOverrides.conf. As above, if you want to customize what is available for an individual course (e.g. one professor may want to try out the new editors and others may not), copy the %showeditors structure to course.conf and then edit it to make available the desired editors. Set up the webwork database WeBWorK uses a single database, called webwork, for all courses. We will create the webwork database now. To do this do the following (before you just copy, paste and hit <Enter> notice that you have to replace database_password with the password you set when editing site.conf above): $ sudo mysql

mysql> CREATE DATABASE webwork;
mysql> CREATE USER 'webworkWrite'@'localhost' IDENTIFIED BY 'database_password';
mysql> GRANT SELECT, INSERT, UPDATE, DELETE, CREATE, ALTER, DROP, LOCK TABLES ON webwork.* TO 'webworkWrite'@'localhost';
mysql> exit
Bye
$git checkout legacy-v2  Configuring Apache WeBWorK ships with an Apache config file that needs to linked into your Apache configuration process. The file is named webwork.apache2.4-config.dist and located in the conf directory. First, copy the file to webwork.apache2-config: $ cd /opt/webwork/webwork2/conf
$cp webwork.apache2.4-config.dist webwork.apache2.4-config  and now link it into your Apache configuration process $ sudo su
# cd /etc/apache2/conf-enabled
# ln -s /opt/webwork/webwork2/conf/webwork.apache2.4-config webwork.conf


The default multi-processing module (MPM) for Apache is the event module but WeBWorK uses the prefork module. So we disable the event module and enable the prefork module.

# a2dismod mpm_event
# a2enmod mpm_prefork


Next we will make a few changes to Apache's default configuration. We need to edit two files. First we will save copies of the original files.

# cd /etc/apache2/
# cp apache2.conf apache2.conf.bak1
# cd /etc/apache2/mods-available/
# cp mpm_prefork.conf mpm_prefork.conf.bak1
# exit


Edit apache2.conf

$cd /etc/apache2/$ sudo nano apache2.conf


Search for the line

Timeout 300


and replace it by

Timeout 1200


Then save the file and quit.

Edit mpm_prefork.conf

$cd /etc/apache2/mods-available/$ sudo nano mpm_prefork.conf


Search for the lines

    MaxRequestWorkers        150
MaxConnectionsPerChild   0


Which occur under <IfModule mpm_prefork_module> and replace them by

# For WeBWorK a rough rule of thumb is 5 MaxRequestWorkers per 1 GB of memory
MaxRequestWorkers        10
MaxConnectionsPerChild   50


where you should set MaxRequestWorkers depending on the amount of memory your server has using the above rule of thumb. Note that for very busy servers, you should observe you memory usage and adjust the above settings as necessary. Also make sure MaxSpareServers is not set too high.

Then save the file and quit.

Finally we copy WeBWorK's icon file favicon.ico to Apache's www directory.

$sudo cp /opt/webwork/webwork2/htdocs/favicon.ico /var/www/html [sudo] password for wwadmin: <wwadmin password>  Now stop and start Apache. We do it this way to make sure the MPM gets changed. $ sudo apache2ctl stop
$sudo apache2ctl start  Test your configuration 1. Test the /webwork2 location by visiting http://yourserver.yourschool.edu/webwork2. You should see the WeBWorK home page with no courses listed. Actually the directory /opt/webwork/courses/ does contain the modelCourse but the modelCourse is not a real course so you will get an error message if you try to log into it. It will be used a as model for setting up other courses. For this reason /opt/webwork/courses/modelCourse/ contains a file named hide_directory and so the modelCourse is not visible. 2. Test the /webwork2_files location by visiting http://yourserver.yourschool.edu/webwork2_files. You should see the "WeBWorK Placeholder Page". 3. You cannot test the /webwork2_course_files location until you have created a course. If Something is Wrong If something is wrong one of the first things to check is that the config files have been edited correctly (e.g. one time a wrapped line in localOverrides.conf caused me problems, another time it was a missing single quote). A quick way to check this is to do a diff between the edited and distributed versions and check that diff reports the changes you made and only those. Another thing is to look at the Apache error log which is located at /var/log/apache2/error.log. $ cd /etc/apache2/
$diff apache2.conf apache2.conf.bak1$ cd /opt/webwork/webwork2/conf/
$diff site.conf site.conf.dist$ diff localOverrides.conf localOverrides.conf.dist
$diff webwork.apache2.4-config webwork.apache2.4-config.dist$ tail /var/log/apache2/error.log


If something is wrong and you fix it, you will have to restart Apache for the changes to take effect

$sudo service apache2 restart [sudo] password for wwadmin: <wwadmin password>  Known Issues Here are the known issues with this release. PGbasicmacros.pl Displaying certain symbols (e.g. {, },<,>,≤, ≥) in the text (not in Math Mode) of a WeBWorK problem fails. An example is Problem 1 in Set 0 in "myTestCourse" (see Test that Things are Working Properly below). For information on this see https://github.com/openwebwork/pg/issues/473. To fix this we need to edit the PGbasicmacros.pl file. $ cd /opt/webwork/pg/macros
$cp PGbasicmacros.pl PGbasicmacros.pl.bak1  $ nano PGbasicmacros.pl


Look for the line

HTML_MathJax     => [ "HTML_dpng", "HTML_tth", "HTML", ],


and replace it by

HTML_MathJax     => [ "HTML_dpng", "HTML", "HTML_tth", ],


Then save the file and Quit.

This may or may not be addressed in a WeBWorK 15 hotfix (see Checking for and Installing Hotfixes).

UTF-8 support

WW 2.15 has some "known issues" with the UTF-8 support that arise in some WeBWorK problems. These issues are discussed in a forum discussion at: https://webwork.maa.org/moodle/mod/forum/discuss.php?d=4795. The discussion there lists pull requests by Nathan Wallach and Davide Cervone that address these issues. If you write or use WeBWorK problems that have issues with UTF-8 characters, look at that discussion. Again these issues may or may not be addressed in a WeBWorK 15 hotfix (see Checking for and Installing Hotfixes).

Course Administration gives information about creating courses. Here we will give explicit instructions for doing this. (You may need to become root or use sudo to execute some of these commands.)

$cd$ newgrp wwdata
$umask 2$ cd /opt/webwork/courses
$/opt/webwork/webwork2/bin/addcourse admin --db-layout=sql_single --users=adminClasslist.lst --professors=admin  Make the course admin and its subdirectories owned by the Apache server. $ sudo chown -R www-data  admin


Updating the OPL

The following assumes you have already installed the OPL. For that see #Install the Open Problem Library below. Now check for updates to the Open Problem Library:

cd /opt/webwork/libraries/webwork-open-problem-library
git remote show origin


This will print several lines of data, but the last line will tell you if your copy is up to date or out of date. If it is out of date, run

git pull origin


Then rerun the OPL-update script. Updating the OPL is pretty much risk free since changes usually involve only a relatively small number of individual problems and the vast majority of problems remain unchanged. Updating the OPL does not require restarting apache2.

$OPL-update  Updating MathJax cd /opt/webwork/MathJax/ git remote show origin  This will print several lines of data, but the second to last line (legacy-v2) will tell you if your copy is up to date or out of date. If it is out of date, run git pull origin  Updating MathJax is usually very safe and does not require restarting apache2. Starting and Stopping Apache and MySQL If you make changes to the system, you will have to restart apache2 before the changes take effect. On rare occasions you may need to restart MySQL. Starting and Stopping Apache You have to run these commands as root. To start or restart (i.e. stop and then start) the apache2 webserver run the command $ sudo apache2ctl graceful


You can also start apache2 by

$sudo apache2ctl start [sudo] password for wwadmin: <wwadmin password>  and restart it with $ sudo apache2ctl restart


restart is less graceful but more powerful than graceful. Sometimes graceful fails to kill all apache2 child processes.

To stop the Apache webserver run the command

$sudo apache2ctl stop [sudo] password for wwadmin: <wwadmin password>  Stopping and then starting Apache should definitely kill all apache2 child processes. You can also start or stop apache2 by using the init.d script apache2. Run $ sudo /etc/init.d/apache2


or equivalently

$sudo service apache2 [sudo] password for wwadmin: <wwadmin password>  and you will get a list of allowed commands (start, stop, restart, etc.). Finally you can use the newer systemctl method, e.g. sudo systemctl start apache2 [sudo] password for wwadmin: <wwadmin password>  sudo systemctl --help  will give you a list of all possible commands, the main ones being start, stop, restart and status. Starting and Stopping MySQL You have to run these commands as root. To start the MySQL server run the command $ sudo /etc/init.d/mysql start


To stop the MySQL server run the command

$sudo /etc/init.d/mysql stop [sudo] password for wwadmin: <wwadmin password>  To restart the MySQL server run the command $ sudo /etc/init.d/mysql restart


Equivalently you can use the command

$sudo service mysql  followed by start, stop, restart, etc. Just running the above alone will display a list of allowed commands. Finally you can use the newer systemctl method, e.g. sudo systemctl start mysql [sudo] password for wwadmin: <wwadmin password>  sudo systemctl --help  will give you a list of all possible commands, the main ones being start, stop, restart and status. Install the WeBWorK Problem Libraries Before we create a real course we will install the WeBWorK Problem Libraries. Fix the modelCourse Unfortunately in the 2.15 distribution someone tried to make setting up the libraries easier but they made mistakes so we have to remove two incorrect links. Do the following: $ cd /opt/webwork/courses/modelCourse/templates/
$rm Contrib$ rm CAPA


We will set up the correct links below.

Install the Open Problem Library

The Open Problem Library consists of both WeBWorK problems and methods for searching and selecting problems. Also it contains as sub libraries many of the other standard libraries. We have to load a database for searching it.

First we put a link to the Open Problem Library in the modelCourse so that when we create courses copying templates from the modelCourse, the OPL will be available.

$cd /opt/webwork/courses/modelCourse/templates/$ ln -s /opt/webwork/libraries/webwork-open-problem-library/OpenProblemLibrary Library


Note that if the above link is not in the templates directory of a course, it will be automatically created when you open the Library Browser. However, if e.g. you haven't yet opened the Library Browser and the link is missing and you import problem sets (as we do in section Test that Things are Working Properly below), you may see errors when viewing problems (you definitely will see errors in the Orientation set).

Next we have to run the OPL-update script which will create the data that the Library Browser uses.

$OPL-update  This has to convert a lot of data for over 37,000 problems so please be patient; it can take a long time. If at some time in the future you want to upgrade the Problem Library, the process is easy, see Updating the OPL. Note that this is something you should do fairly often so that your library is up to date with new problems and bug fixes for old ones. Initialize the OPL local statistics database table Run the command $ update-OPL-statistics.pl


to create a table to hold the OPL local statistics. See https://webwork.maa.org/wiki/OPL_Problem_Statistics for information.

If this table does not exist, when you try to view problems in the Library Browser, you will see the following error message:

"couldn't find the opl local statistics table"

The Contrib directory contains contributions to the OPL that may not have been formally accepted into the main collection of OPL problems. This may be because the contribution is recent and has not yet been reviewed or because the problems are not properly tagged as is the case with the CAPA problem collection. It also contains the original versions of problems that have been accepted into the OPL (the idea being that the original authors maintain control of problems in the Contrib directory but problems in the OPL are controled by the editors). This step creates a button in the Library Browser which gives a direct link to the Contrib directory.

Put a link to the Contrib directory in the modelCourse so that when we create courses copying templates from the modelCourse, the Contrib directory will be available.

$cd /opt/webwork/courses/modelCourse/templates/$ ln -s /opt/webwork/libraries/webwork-open-problem-library/Contrib Contrib


If you just want to do this for individual courses, not all courses then don't put the link in the modelCourse above, and then do the following. Note we don't have any individual courses yet. But for example after creating myTestCourse below, to set up access to the Contrib directory from myTestCourse, do the following

$cd /opt/webwork/courses/myTestCourse/templates/$ sudo $ln -s /opt/webwork/libraries/webwork-open-problem-library/Contrib Contrib [sudo] password for wwadmin: <wwadmin password>  Set up the Rochester and Union Libraries This step is optional. It creates buttons in the Library Browser which give direct links to the Rochester and Union libraries. If you don't do this, you can find these libraries and others under the OPL Directory button. First we need to edit localOverrides.conf one last time $ cd /opt/webwork/webwork2/conf
$nano localOverrides.conf  Search for courseFiles{problibs} and scroll down several lines to the lines # rochesterLibrary => "Rochester", # unionLibrary => "Union",  Uncomment these lines (i.e. remove the #) so they become  rochesterLibrary => "Rochester", unionLibrary => "Union",  Then save the file and quit. We next put links to the Rochester and Union Libraries in the modelCourse so that when we create courses copying templates from the modelCourse, these libraries will be available. Skip this step if you usually only want to use Open Problem Library. Note that the Rochester, Union and other libraries are contained in the OPen Problem Library and are accessible from there under the OPL Directory button in the Library Browser. This step simply creates buttons in the Library Browser so that you can access the Rochester and Union libraries directly. $ cd /opt/webwork/courses/modelCourse/templates/
$ln -s /opt/webwork/libraries/webwork-open-problem-library/OpenProblemLibrary/Union unionLibrary$ ln -s /opt/webwork/libraries/webwork-open-problem-library/OpenProblemLibrary/Rochester rochesterLibrary


If you want to put another library into the modelCourse, just do the analogous thing. If you just want the additional library in a particular course, add the link in the templates directory of that course. If you look in the directory /opt/webwork/libraries/webwork-open-problem-library/OpenProblemLibrary/ you might find other libraries that are not yet listed in global.conf and these can be added in the same way as the Rochester and Union libraries. Finally if you add a library with non standard symbols in the name (e.g. uva-statLibrary) you have to use single quotes when adding it to global.conf, e.g.
'uva-statLibrary' => "UVA-Stat",
It's easier to just avoid such names.

Install and Set Up the CAPA Library

This step is optional. It installs and sets up the CAPA Library, which is a library of physics problems.

Note: We are in the process of integrating the CAPA problems into the main distribution so some of the details below may change.

Move the CAPA graphics files to the required locations and set the group.

$cd /opt/webwork/libraries/webwork-open-problem-library/Contrib/CAPA$ mv CAPA_Graphics /opt/webwork/webwork2/htdocs/
$sudo chgrp -R wwdata /opt/webwork/webwork2/htdocs/CAPA_Graphics [sudo] password for wwadmin: <wwadmin password>$ sudo chmod -R g+w /opt/webwork/webwork2/htdocs/CAPA_Graphics


We need to edit localOverrides.conf again

$cd /opt/webwork/webwork2/conf$ nano localOverrides.conf


Now search for the lines

$pg{specialPGEnvironmentVars}{CAPA_Tools} = "$courseDirs{templates}/Contrib/CAPA/macros/CAPA_Tools/",
$pg{specialPGEnvironmentVars}{CAPA_MCTools} = "$courseDirs{templates}/Contrib/CAPA/macros/CAPA_MCTools/",
$pg{specialPGEnvironmentVars}{CAPA_GraphicsDirectory} = "$courseDirs{templates}/Contrib/CAPA/CAPA_Graphics/",
push @{$pg{directories}{macrosPath}}, "$courseDirs{templates}/Contrib/CAPA/macros/CAPA_Tools",
"$courseDirs{templates}/Contrib/CAPA/macros/CAPA_MCTools";  and replace these six lines by ################################################################################ #Locations of CAPA resources. (Only necessary if you need to use converted CAPA problems.) ################################################################################$pg{specialPGEnvironmentVars}{CAPA_Tools}             = "$courseDirs{templates}/Contrib/CAPA/macros/CAPA_Tools/",$pg{specialPGEnvironmentVars}{CAPA_MCTools}           = "$courseDirs{templates}/Contrib/CAPA/macros/CAPA_MCTools/",$pg{specialPGEnvironmentVars}{CAPA_GraphicsDirectory} = "$webworkDirs{htdocs}/CAPA_Graphics/";$pg{specialPGEnvironmentVars}{CAPA_Graphics_URL}      = "$webworkURLs{htdocs}/CAPA_Graphics/";  Then save the file and Quit. There is one final step that is needed. We have to put a link in the templates directory of every course that needs access to the CAPA Library. If you want to have every course you create have access to the CAPA Library (unlikely unless you are in a physics department) put the link in the modelCourse. $ cd /opt/webwork/courses/modelCourse/templates/
$ln -s /opt/webwork/libraries/webwork-open-problem-library/Contrib/CAPA/ capaLibrary  More likely if you just want to do this for individual courses, not all courses then don't put the link in the modelCourse above, and then do the following. We don't have any yet. But for example after creating myTestCourse below, to set up access to the CAPA Library from myTestCourse, do the following $ cd /opt/webwork/courses/myTestCourse/templates/
$sudo ln -s /opt/webwork/libraries/webwork-open-problem-library/Contrib/CAPA/ capaLibrary [sudo] password for wwadmin: <wwadmin password>  and do the analogous thing for every course that needs access to the CAPA Library. Then to gain access to the CAPA Library from the course, simply go to the Library Browser and click on the CAPA button. Create Your First Actual Course Since we have edited site.conf and localOverrides.conf and these are critical files, it would be a good idea to run $ cd /opt/webwork/webwork2/conf
$diff site.conf site.conf.dist$ diff localOverrides.conf localOverrides.conf.dist


and check that you haven't made any mistakes (e.g. by introducing an inadvertent line break, etc.). If there are any mistakes, correct them. Remember that any time you change either of these files you must restart the Apache webserver in order for these changes to take effect.

Since we have edited global.conf and haven't restarted Apache we do so now.

$sudo apache2ctl graceful password:<wwadmin password>  Now log into the admin course ( http://yourserver.yourschool.edu/webwork2/admin or, if you have not hidden the admin course, click on Course Administration on WeBWorK's home page http://yourserver.yourschool.edu/webwork2/ ) as yourself or admin and 1. click on Add Course 2. For Course ID enter myTestCourse 3. For Course Title enter My Test Course 4. Enter your institution 5. Leave Add WeBWorK administrators to new course checked 6. Add an additional instructor if you wish 7. Copy templates from: modelCourse (the default action) 8. Click on Add Course 9. Click Log into myTestCourse and log in either as admin or yourself (if you added yourself as an additional instructor above). At some point you will probably want to "hide" myTestCourse from general view but you already know how to do that. Test that Things are Working Properly We will test out a few important parts of WeBWorK. If you run into problems, you should look at the Apache error log which is located at /var/log/apache2/error.log. Click on Hmwk Sets Editor on the Main Menu. Then select Import, select setDemo.def from the from drop down list and select all current users from the assigning this set to drop down list. Then hit Take Action! Now click on Homework Sets on the Main Menu and click on Demo. Then look at the problems. Mathematical equations should be typeset. If not, edit the file Constants.pm in the directory /opt/webwork/webwork2/lib/WeBWorK. Change the line $WeBWorK::PG::ImageGenerator::PreserveTempFiles = 0; to ...::PreserveTempFiles = 1;. Then restart Apache and view the first couple problems or some new ones. Then look in the directory /opt/webwork/webwork2/tmp/. cd to one of the ImageGenerator.../tmp/ directories and look at the error and log files there. When you fix the problem remember to edit ...::PreserveTempFiles = 1; back to 0 and restart Apache or you will be saving a lot of unnecessary files. Another useful trick is to try downloading a hard copy of an assignment and then (assuming there are errors) looking at the various log files that are linked to on the output page.

Note that since you probably used "admin" to login you have Permission Level "admin" so things look more complicated than if you were just a regular professor with Permission Level "professor". In particular any time you view a problem, you have the option to "Edit tags" which you probably do not want to do. If these bothers you, login as a professor (see below).

Continue looking at problems to see if everything is working properly.

Next click on Problem List to bring back the Problem List Page and click on Download PDF .... The page is a little complicated because you are an admin so you see the professor view (students see a very simple page) but you can just scroll to the bottom and click on Generate hardcopy for selected users and selected sets. If you get an error (you shouldn't) just click Download Hardcopy to get what was generated. Also you can see links to various informational files that are available if you run into problems (normally these files are removed if there are no errors). If you want to preserve these tmp files, set $WeBWorK::ContentGenerator::Hardcopy::PreserveTempFiles to 1 in the file Constants.pm in the directory /opt/webwork/webwork2/lib/WeBWorK and then restart Apache. Remember to set this back to 0 after debugging. Another thing to do is to use Email on the Main Menu. Again this page is a little complicated because you can do a lot of things with it (including mail merge) but at this point just select yourself in the list to the right and hit Send Email at the bottom. You should receive two emails. One is the message you just sent and the other is an email with subject "WeBWorK email sent" giving information on your mailing. As a final test click on Library Browser on the Main Menu. Click Open Problem Library  (actually it should already be selected so it will be greyed out) and select a Subject, Chapter and Section and then hit View Problems. The first 20 of your selected problems will be displayed. You can also test that you can access any additional Problem Libraries that you installed. Note that if you just happen to view some statistics problems that require the R server, you will get errors (since we have not installed R --- at least not yet). Note: if when using the Library Browser you find that after selecting a Subject, you can not select a Chapter (and usually you will see an error from setmaker.js: /webwork2/instructorXMLHandler: e.g. Timeout or Forbidden), the most common reason is that$server_root_url has not been set "correctly" in the /opt/webwork/webwork2/conf/site.conf file. A common error is to use forget http:// or to use http:// when you should use https:// or to just have the wrong domain name or ip address. But sometimes "localhost" will work when the correct ip address or url will not, for example in situations when you are using a virtual machine as your host (e.g. on AWS).

If all the above tests work, you can be pretty confident that WeBWorK is working properly.

Go back to Hmwk Sets Editor on the Main Menu. Then select Import, select setOrientation.def from the from drop down list and select all current users from the assigning this set to drop down list. Then hit Take Action!. Then go through the Orientation problems. This is a good first set to use for introducing students to WeBWorK. Also you can Import set0.def and setMAAtutorial.def and look through those problems.

If you are new to WeBWorK or even if you are a pro, you should probably add a regular student to myTestCourse and log in as that student to see what the student interface looks like. It's much simpler than the professor or admin interfaces. Also for the same reason you should probably add professor to myTestCourse and log in as that professor to see what the professor interface looks like.

Click on Classlist Editor on the Main Menu. Then select Add, for "Add how many students?" pick 2 and hit Take Action!. Then add two students, say

• Last Name "Smith", First Name "Jane" , Student ID "jsmith" and Login Name "jsmith"
• Last Name "A", First Name "Prof" , Student ID "profa" and Login Name "profa"

and make sure to select all sets to assign them to both of them. Then hit Add Students.

Now we make "profa" a professor. Again click on Classlist Editor on the Main Menu. The fast way is to click on the "pencil" by "profa", set the "Permission Level" to "professor" and hit Take Action!.

Jane Smith's initial password will be her Student ID jsmith. Now login as Jane Smith and play around a little.

Prof A's initial password will be her Student ID profa. You should also login as Prof A and play around a little.

You can also add "practice users" and login as one of those to see how you can grant partial access to your course for practice or other reasons (e.g. auditors who don't want to be graded). To do this Click on Classlist Editor, select Import, select the "demoCourse.lst" (the only classlist available at this point) and hit Take Action!.

Optional Configurations

• Optional A stores WeBWorK's "temporary" files in a separate partition (or directory).
• Optional B installs and configures a lightweight webserver to serve static files.
• Optional C configures Apache so that access to WeBWorK will be through SSL.
• Optional D installs and configures the R server for statistical computing.
• Optional E sets up log rotation for WeBWorK's timing log.

Implement Optional A (wwtmp)

Now is the time to implement Optional A if you choose to do so. Actually you can do this at any time and your active courses will continue to function seemingly without change. The only change behind the scenes will be that temporary files will be stored in a different location. Note that if you want to use this option but did not create wwtmp as a separate partition, you first have to create the directory /var/www/html/wwtmp with the command

sudo mkdir /var/www/html/wwtmp


All of WeBWorK's "temporary" files will be stored under /var/www/html/wwtmp. These are mostly small files such as png images of equations, pdf files, etc. that may be reused but if they are not present (e.g. if they get deleted) they will be seamlessly regenerated on the fly. There is no reason to back up such files and having them in a separate partition or directory means that it is easier and faster to back up other partitions and skip backing up unnecessary files. Even if you do not want to set up a separate partition for this, it is very convenient to at least set up a separate directory for these temporary files and we recommend that yo do so.

First we set the group and permissions for the wwtmp directory

$sudo su [sudo] password for wwadmin: <wwadmin password> # cd /var/www/html # chown wwadmin wwtmp # chgrp wwdata wwtmp # chmod ug+w wwtmp # chmod g+s wwtmp # exit$


Next we have to edit localOverrides.conf so that WeBWorK uses the new wwtmp directory. Since we have a working WeBWorK system, first we make a backup copy of localOverrides.conf.

$cd /opt/webwork/webwork2/conf$ cp localOverrides.conf localOverrides.conf.bak1
$nano localOverrides.conf  Now edit localOverrides.conf. Search for the line # Directory for temporary files  and under that find the line # To implement, uncomment the following 6 lines:  Then uncomment (i.e. remove the # symbol) from the 6 lines #$webworkDirs{htdocs_temp}  =  '/var/www/html/wwtmp';
#$webworkURLs{htdocs_temp} = '/wwtmp'; #$webworkDirs{equationCache} = "$webworkDirs{htdocs_temp}/equations"; #$webworkURLs{equationCache} = "$webworkURLs{htdocs_temp}/equations"; #$courseDirs{html_temp}  =  "/var/www/html/wwtmp/$courseName"; #$courseURLs{html_temp}   = "/wwtmp/$courseName";  IMPORTANT NOTE. Apache 2.4 has changed the default location of the document root from /var/www to /var/www/html and we are in the process of updating the file localOverrides.conf.dist to reflect this. If your copy of this file does not have /html in the two lines above, then in addition to uncommenting the 6 lines you also have to add /html in two places. Then save the file and quit. If you look at the wwtmp directory you will find it empty but after you restart apache and then access some WeBWorK problems, you will find temporary directories and files in wwtmp. Remember your have to restart apache for these changes to take effect. Using Cron Jobs to remove temporary files It is a good idea to clean out temporary files on a regular automatic schedule. Also pdf copies of downloaded problem sets are saved in a temporary directory (wwtmp/.../hardcopy) so that they can be downloaded from the web. But after the download, the pdf file remains and is visible from the web if one knows the URL. For this reason we recommend deleting all such files that are over one hour old. Similarly we recommend deleting all png, gif, and html links under wwtmp that are over 30 days old. And finally every week we recommend deleting all equation images that are over 14 days old. The following cron jobs will accomplish this. The first is run every 30 minutes, the next three twice a month and the last one weekly on Sunday morning. These cron jobs should be run as root. We use crontab to edit the crontab file: $ sudo su
# crontab -e


Now add the following lines at the end of the file

WEBWORK_ROOT=/opt/webwork/webwork2
*/30 * * * *  find /var/www/html/wwtmp/*/hardcopy/*  -mmin +60  -name "*" -delete
5 5 1,15 * *  find /var/www/html/wwtmp/*/gif/  -mtime +30  -name "*" -delete
5 5 2,16 * *  find /var/www/html/wwtmp/*/png/  -mtime +30  -name "*" -delete
5 5 3,17 * *  find /var/www/html/wwtmp/*/html/  -mtime +30  -name "*" -delete
4 5 * * 0 /opt/webwork/webwork2/bin/remove_stale_images --delete --days 14


and save the file and quit

# exit
$ Implement Optional B (lighttpd) As is the case for Optional A you can implement Optional B at any time and your active courses will continue to function seemingly without change. The only change behind the scenes will be that static images, pages and MathJax will be served by a light weight web server. Install and Configure lighttpd First we install the light weight webserver lighttpd Run the command $ sudo apt install lighttpd


Now we configure lighttpd. First let's make a backup of the configuration file.

$cd /etc/lighttpd$ sudo cp lighttpd.conf lighttpd.conf.bak1


Now edit lighttpd.conf.

$sudo nano lighttpd.conf  Apache2 is listening on port 80 so we need an alternate port for lighttp to listen to. Standard alternate ports for this are usually 81, 8000, or 8080. 8080 is the only port that is listed as an official alternate at http://en.wikipedia.org/wiki/List_of_TCP_and_UDP_port_numbers . Note that in rare cases an institution may block httpd requests to port 8080. If any of your students report that they can not see graphics, they will have to request that access to requests to port 8080 be allowed. Find the line server.port = 80  and replace it by server.port = 8080  Now we make a few more changed that allow MathJax to run under lighttpd. At the top of the file in the section  server.modules = , under the line  "mod_redirect",  add the line "mod_setenv",  Then below the closing parentheses and above the line server.document-root = "/var/www/html"  add the lines setenv.add-response-header = ( "Access-Control-Allow-Origin" => "*" ) alias.url = ( "/webwork2_files" => "/opt/webwork/webwork2/htdocs/" )  Then save the file and quit. Now restart lighttp $sudo /etc/init.d/lighttpd restart



Note that you can just run /etc/init.d/lighttpd to get a list of all options.

Test lighttpd

First run the command

 $sudo lsof -i -P -n | grep LISTEN  and check that lighttpd is listening on port 8080 Now test your server by connecting to your server from a browser on a remote machine ("http://yourserver.yourschool.edu:8080/"). You should see the Apache2 Ubuntu Default Page but without a graphic for the Ubuntu Logo indicating that lighttp is running . Test the /webwork2_files location by visiting "http://yourserver.yourschool.edu:8080/webwork2_files. You should see the "WeBWorK Placeholder Page". Configure WeBWorK to Take Advantage of Lighttp First let's make a backup copy of localOverrides.conf  so that we can easily back out of these changes if necessary. $ cd /opt/webwork/webwork2/conf
$cp localOverrides.conf localOverrides.conf.bak2  Now edit localOverrides.conf. Note that while Optional B is independent of Optional A, we assume most people implementing Optional B will have already implemented Optional A. Therefore we give instructions for editing localOverrides.conf assuming that Optional A has been implemented. If this is not the case, modify the instructions below accordingly. Also replace yourserver.yourschool.edu with the correct address. $ nano localOverrides.conf


Find the line

$webworkURLs{htdocs_temp} = '/wwtmp'  and replace it by #$webworkURLs{htdocs_temp}   = '/wwtmp';
$webworkURLs{htdocs_temp} = 'http://yourserver.yourschool.edu:8080/wwtmp';  Find the line $courseURLs{html_temp}   = "/wwtmp/$courseName";  and replace it by #$courseURLs{html_temp}   = "/wwtmp/$courseName";$courseURLs{html_temp}   = "http://yourserver.yourschool.edu:8080/wwtmp/$courseName";  Under this line add the lines # Location of MathJax script, used for the MathJax display mode.$webworkURLs{MathJax}       = 'http://yourserver.yourschool.edu:8080/webwork2_files/mathjax/MathJax.js?config=TeX-MML-AM_HTMLorMML-full';


Then save the file and quit.

Now restart apache and lighttp.

$sudo apache2ctl graceful password:<wwadmin password>$ sudo /etc/init.d/lighttpd restart


Test that Everything is Working Properly

To test things go to your test course http://yourserver.yourschool.edu/webwork2/myTestCourse/. Log into your course and view a problem with a graphic image (e.g. Problem 2 of the Demo set. Since you have Admin or Prof privileges, you can view sets which are not yet open). Right click on the image and click on Properties (or whatever is appropriate on your browser, e.g. copy image location) and check that the image is being served from port 8080 (something like http://yourserver.yourschool.edu:8080/wwtmp/myTestCourse//gif/...). To test that MathJax is using lighttpd, view a problem with some typeset equations. Right click on the equation and you should see the MathJax menu which confirms MathJax is being used. Next look at the source code for the page (e..g. right click most browsers) and in the source, search for mathjax.js. You should see that this is being loaded from port 8080.

Disabling Optional B (lighttpd) for a single course

If your server is hosting courses from different institutions, you may find that you need to disable using lighttpd for a single course, e.g. because the institution blocks access to port 8080. Actually since port 8080 is an "official" alternate httpd port, you should first try to get the network administrators at the institution to grant access to port 8080. Failing this, you can disable using lighttpd and instead use apache for a single course by adding the following lines to the end of the course's course.conf file which is located in the top level directory for the course (usually /opt/webwork/courses/course_name). Note that using apache instead of lighttpd will put a slightly larger load on the server. Also note that you need "admin" level privileges to edit the course.conf file. Instructors with only "professor" level privileges can not do this from within WeBWorK.

# Do not use lightppd (port 8080) for this course
$webworkURLs{htdocs_temp} = '/wwtmp';$courseURLs{html_temp}   = "/wwtmp/$courseName";$webworkURLs{equationCache} = "$webworkURLs{htdocs_temp}/equations";$webworkURLs{MathJax}       = "$webworkURLs{htdocs}/mathjax/MathJax.js?config=TeX-MML-AM_HTMLorMML-full";  Implement Optional C (SSL) Optional C configures apache so that access to WeBWorK will be through an encrypted connection (SSL) with an https: URL. Note that TLS is the successor protocol to SSL and is used everywhere. So that when we and others use the more common acronym SSL, we really are talking about TLS. If you have implemented Optional B, the non encrypted lighttpd server will be used for images, MathJax, etc. so we might have to set up the lighttpd server to run under SSL (see below). First we have to obtain an official SSL certificate or generate a self-signed one which can be used for testing purposes. If students will be using your server, it is better to obtain an official SSL certificate since using a self-signed certificate with cause warning messages to be displayed when students connect to the server. Obtain an Official SSL Certificate Talk to the networking group on your campus about obtaining an official certificate. You can find information on the net about certificate providers, e.g. http://www.sslshopper.com/certificate-authority-reviews.html . The certificate goes in /etc/ssl/certs/ and the key file goes in /etc/ssl/private/. Use a self-signed Certificate When we installed the package openssh-server a self-signed certificate was automatically created. The certificate is stored at /etc/ssl/certs/ssl-cert-snakeoil.pem  and the private key at /etc/ssl/private/ssl-cert-snakeoil.key  This certificate and key may be regenerated manually with the following command (needs root privileges ie sudo): $ sudo make-ssl-cert generate-default-snakeoil --force-overwrite


which you might want to do since I believe the original certificate is only valid for 365 days.

In the instructions below we will use the provided self-signed certificate (ssl-cert-snakeoil.pem and ssl-cert-snakeoil.key) but you can also generate your own using openssh following instructions on the web. As we said above, for a production server, you should really use an official certificate.

Set up Apache to use SSL

First we enable the mod_ssl module

$sudo a2enmod ssl [sudo] password for wwadmin: <wwadmin password>  Now we have to configure Apache to use SSL. $ cd /etc/apache2/sites-available/
$sudo cp default-ssl.conf default-ssl.conf.bak1 [sudo] password for wwadmin: <wwadmin password>$ sudo nano default-ssl.conf


Our self-signed certificate and key files are named ssl-cert-snakeoil.pem and ssl-cert-snakeoil.key. If you are using official files, put their names in the lines below. You will also have to edit the Certificate Chain, Certificate Authority and possibly other items. Instructions for doing so are beyond the scope of this document. You might look at https://www.digitalocean.com/community/tutorials/how-to-secure-apache-with-let-s-encrypt-on-ubuntu-18-04

Search for the lines

SSLCertificateFile      /etc/ssl/certs/ssl-cert-snakeoil.pem
SSLCertificateKeyFile /etc/ssl/private/ssl-cert-snakeoil.key


and replace ssl-cert-snakeoil.pem and ssl-cert-snakeoil.key by the names of your certificate. Since in these instructions we are using ssl-cert-snakeoil.pem and ssl-cert-snakeoil.key, we can just leave the lines as is.

Then save the file and quit. And enable default-ssl.conf

$sudo a2ensite default-ssl.conf [sudo] password for wwadmin: <wwadmin password>  Finally we restart Apache $ sudo systemctl restart apache2


and test things. Connect to https://yourserver.yourschool.edu/webwork2/myTestCourse You will be asked to accept the certificate. After you do so things should work just as before except that the connection will be via https except for images, MathJax, etc. if you using lighttp. In that case, viewing math expressions in image mode should work fine but it is possible using MathJax mode will fail. We will talk about this below.

Redirect http requests to https

Assuming that everything is working, the last thing we do is set things up so that requests to http://yourserver.yourschool.edu/webwork2/ are automatically redirected to https://yourserver.yourschool.edu/webwork2/.

$cd /etc/apache2/sites-available$ sudo cp 000-default.conf 000-default.conf.bak1
$sudo nano 000-default.conf  In the <VirtualHost *:80>  section just under the line DocumentRoot /var/www/html  add the line Redirect permanent /webwork2 https://yourserver.yourschool.edu/webwork2  where of course you should edit yourserver.yourschool.edu appropriately. Then save the file and quit. Restart Apache $ sudo service apache2 restart


and try connecting to http://yourserver.yourschool.edu/webwork2/. The real connection should be through https://yourserver.yourschool.edu/webwork2/.

Check site.conf

This is important. We have to check the value of $server_root_url in webwork2/conf/site.conf to make sure it uses https (so it should read e.g. https://yourserver.yourschool.edu). If you don't do this, your may run into strange problems in the Library Browser. To check, open the Library Browser, select a subject and then click on the chapter area. If you see a drop down list of chapters, things are working properly. Now we check and edit if necessary site.conf cd /opt/webwork/webwork2/conf nano site.conf  In the line $server_root_url = ...


replace http by https if necessary. Then save the file, quit and restart apache

sudo service apache2 restart


Configure lighttpd to use SSL

We are using lighttpd for images, pdf's, MathJax, etc. MathJax may fail (so mathematical expressions will not be displayed) if you do not configure lighttpd to use SSL. So let's configure lighttpd to use SSL.

For lighttpd you need to concatenate the key file and the certificate file into a single pem file by running the following command. Obviously you should use the real name of your certificate and key files, not "apache.key" and "apache.crt".

$sudo su [sudo] password for wwadmin: <wwadmin password> # cd /etc/ssl/private # cat apache.key ../certs/apache.crt > apache.pem # chmod 640 apache.pem # exit$


If you are using official certificates, you will also have to edit the Certificate Chain, Certificate Authority and possibly other items. Instructions for doing so are beyond the scope of this document. See e.g. http://redmine.lighttpd.net/projects/lighttpd/wiki/Docs_SSL and/ or https://www.vultr.com/docs/setup-let-s-encrypt-with-lighttpd-on-ubuntu-16-04

Since we already are using the files ssl-cert-snakeoil.pem and ssl-cert-snakeoil.key for apache, we will also use them for lighttpd

$sudo cat /etc/ssl/private/ssl-cert-snakeoil.key /etc/ssl/certs/ssl-cert-snakeoil.pem >/etc/lighttpd/lighttpd.pem [sudo] password for wwadmin: <wwadmin password>  Now edit lighttpd.conf. $ cd /etc/lighttpd
$sudo cp lighttpd.conf lighttpd.conf.bak2 [sudo] password for wwadmin: <wwadmin password>$ sudo nano lighttpd.conf


First in the list of server.modules at the top of the file, under the last module "mod_setenv",, add the module

"mod_openssl",


Next under the line

server.port                 = 8080


$SERVER["socket"] == ":8443" { ssl.engine = "enable" ssl.pemfile = "/etc/lighttpd/lighttpd.pem" }  Then save the file and quit. And restart lighttpd and check it's status $ sudo systemctl restart lighttpd
$sudo systemctl status lighttpd  Sometimes when Lighttpd would not restart (or start) I just rebooted the server and everything worked. If you have difficulty you can run the command $ sudo lsof -i -P -n | grep LISTEN


which will list all the ports your server is listening to. In the list you should see something similar to

lighttpd  519        www-data    5u  IPv4  18610      0t0  TCP *:8443 (LISTEN)


Edit localOverrides.conf

$cd /opt/webwork/webwork2/conf$ cp localOverrides.conf localOverrides.conf.bak4
$nano localOverrides.conf  and replace all occurrences of 8080 by 8443 (one occurrence in three different lines). Then in these same three lines replace http by https . Obviously, you have to edit the three active lines and can ignore the commented out lines. Then save the file and quit. And restart apache $ sudo service apache2 restart


and test that all is well by viewing a page with math expressions using MathJax mode.

Implement Optional D (Rserve)

Here we will install and configure the R server for statistical computing. Many statistics and probability problems (mostly contributed by faculty at the University of British Columbia) require this.

See the documentation (from which we have borrowed heavily) R_in_WeBWorK for more detailed information.

Install the R server

$sudo apt install r-cran-rserve [sudo] password for wwadmin: <wwadmin password>  and check that R is running $ R


You should see something very similar to

R version 3.6.3 (2020-02-29) -- "Holding the Windsock"
Copyright (C) 2020 The R Foundation for Statistical Computing
...
Type 'q()' to quit R.

>


and then quit

> q()
Save workspace image? [y/n/c]: n


Configure Webwork with the location of the R server

Edit the localOverrides.conf file

$cd /opt/webwork/webwork2/conf$ nano localOverrides.conf



Search for the line

$pg{specialPGEnvironmentVars}{DragMath}  and under that line add the following ################################################################################ #location of the R server ################################################################################$pg{specialPGEnvironmentVars}{Rserve} = {host => 'localhost'};


Then save the file and Quit.

Set up Rserve to run as a daemon at system boot

This is somewhat complicated. I'm basically following instructions from https://github.com/geordielad/rserve-systemd-unit

First we create a user with limited privileges to run rserve.

$sudo adduser rserveuser --system --shell=/bin/false --no-create-home --disabled-password --disabled-login  Next we create a file to automatically start Rserve $ cd
$nano rserve.service  Now copy and paste the following code into nano [Unit] Description=Rserve [Service] Type=forking PIDFile=/var/run/rserve/rserve.pid # Define runtime directory RuntimeDirectory=rserve RuntimeDirectoryMode=750 User=rserveuser Group=nogroup ExecStart=/usr/bin/R CMD Rserve --quiet --vanilla --RS-pidfile /var/run/rserve/rserve.pid [Install] WantedBy=multi-user.target  and then save the file and quit. Change the owner to root, move the file to the proper location and register it. $ sudo chown root:root rserve.service
$sudo mv rserve.service /etc/systemd/system$ sudo systemctl daemon-reload


Finally we enable the service to start when the system starts

$sudo systemctl enable rserve.service  Test that things are working properly First start Rserve manually $ sudo systemctl start rserve


and run

$sudo lsof -i -P -n | grep LISTEN  and you should see that Rserve is listening on port 6311 Next test that WeBWorK problems using R run properly. In WeBWorK, click on Library Browser on the Main Menu. Then click Open Problem Library  (actually it should already be selected so it will be greyed out). Now select "Statistics" as Subject, "Bayesian inference" as Chapter and "Posterior distribution" as Section and then hit View Problems. If the problems display with no error messages, all should be well. To be totally sure, click on the "eye" (Try it) in the upper right corner and test the problem. If there are no error messages, congratulate yourself. Everything works. Finally reboot your server (to test that the Rserve starts automatically on boot up) and run $ sudo systemctl status rserve.service


to check that all is well and just to be redundant

$sudo lsof -i -P -n | grep LISTEN  again. Now we are sure that Rserve automatically starts when the system is booted. Implement Optional E (Logrotate) Here we will set up log rotation for WeBWorK's timing log which can grow huge especially if you have a large number of users. The timing.log gives timing information on every action in WeBWorK (see WeBWorK_performance) and can be used to assess the performance of your server. Create a WeBWorK configuration file for Logrotate Logrotate’s configuration files are located in /etc/logrotate.d. Now we create one for WeBWorK. $ sudo nano /etc/logrotate.d/webwork


Now copy and paste the following code into nano

/opt/webwork/webwork2/logs/timing.log {
su www-data wwdata
weekly
missingok
rotate 15
compress
delaycompress
notifempty
}


and then save the file and quit. The backup will be done weekly and we are keeping 15 weeks (approximately one semester) worth of logs. The most recent backup will not be compressed but all older ones will be. We are not explicitly creating a new timing.log file because WeBWorK will automatically create one if it's missing. If you want to check that it works you can run logrotate with the configuration file (make sure you have done something with WeBWorK so that the timing.log is not empty):

sudo logrotate -f /etc/logrotate.d/webwork


Then if you look in /opt/webwork/webwork2/logs, you should see

timing.log.1


You will not see a timing.log file unless someone has been doing something with WeBWorK. If there is no timing.log file do something yourself (e.g. login to a course) and then you should see

timing.log
timing.log.1


Now run logrotate again

sudo logrotate -f /etc/logrotate.d/webwork


and do something in WeBWorK and you should see

timing.log
timing.log.1
timing.log.2.gz


Here is what is going on. timing.log.1 was renamed to timing.log.2 and compressed giving timing.log.2.gz. timing.log was renamed to timing.log.1. And the something you (or someone else) did in WeBWorK created a new timing.log file. A similar procedure will happen weekly until there are a total of 15 backups after which the oldest one will be deleted.

Where to go From Here

You should play around with myTestCourse e.g. click on Library Browser and browse the Problem Library.