static-bugzilla

What is this?

This is a hacky little piece of code I wrote to convert my ancient and abandoned Bugzilla installation into a static, read-only archive.

This let me decommission the Bugzilla instance but leave the data publically available, keeping the most important URLs continuing to function. As an added benefit, the data serves faster and the server no longer needs to keep Bugzilla, MySQL, jobqueue.pl, etc, running.

It might work for you.

It pulls all the data it needs from the public website, so you can do it without direct access to the database. This means you could probably use it to archive someone else's Bugzilla install that you don't control, given time and bandwidth to pull everything.

It is intended to make the individual bug pages and attachments available on their original URLs; everything else goes to a 404 page.

How to use this.

Clone this repository somewhere. Make sure perl and the curl command line tool are installed. I ran this on Linux, it probably works on macOS, Windows will probably need some small effort to set up.
If you intend to replace Bugzilla instead of just keep a static archive for yourself, then on the Bugzilla webserver, you will need a way to run PHP scripts. You will no longer need Perl, MySQL, or cgi-bin support.
Edit the top of make_bugzilla_static_readonly.pl. Look for the block that says # EDIT THESE LINES TO FIT YOUR BUGZILLA INSTALL! ... read each line and change it to match your target installation.
Run make_bugzilla_static_readonly.pl from the directory it is installed in. It will create directories in the same place as it works. Make sure you can access the Bugzilla installation over the network from this computer.
The script will begin pulling in data it needs from Bugzilla. If the script is interrupted, it will pick up where it left off next time. It doesn't move files into place until they are ready, so a power outage shouldn't leave data half-prepared.
Downloading may take significant time. Running this script directly on the webserver itself (so no bandwidth delays) still took about four hours to download 6678 bugs and 3639 attachments, as each HTTP request to Bugzilla can take a few seconds. I shudder at running this against mozilla.org's bugtracker which is dangerously close to 2 million bugs at the time of this writing.
After download, the script will start generating static HTML pages for each bug. This is pretty fast; my laptop chewed through 6678 bugs in 38 seconds, but like the downloads, it is able to pick up where it left off if interrupted.

If you're replacing Bugzilla:

Edit index.html and replace noreply@example.com with your email address (or change the text in some other way).
After the script runs successfully, shut down Bugzilla on your server and move the install out of the way. Copy this directory in its place, so that show_bug.php will serve from the Bugzilla instance's root URL. The attachments and bugs-html directories that the script generated should be here, too.
Change your webserver install so that show_bug.cgi requests go to show_bug.php and attachment.cgi requests go to attachment.php...and that all of these requests run as PHP pages. Note that many Bugzilla installations have a second domain for attachments (for some security reason I never understood), so attachment.cgi might run from, say, bugzilla.example.com and also bugzilla-attachments.example.com, depending on your site's config.)
Make sure the webserver serves index.html for all 404 responses, which explains that Bugzilla is gone and most of its URLs are no longer available. Feel free to edit this index.html as appropriate.
Make sure the webserver serves images/favicon.ico and has the correct mimetype.
There is an Apache2 .htaccess file to block off files that shouldn't be accessed directly on the web; you might have to deal with yourself this if not using Apache.

My Apache virtual host config looks something like this:

ServerAdmin bugmaster@example.com
DocumentRoot "/webspace/bugzilla.example.com"
ServerName bugzilla.example.com
ServerAlias bugs.example.com
ServerAlias bugzilla-attachments.example.com

ErrorDocument 404 /index.html
Alias /show_bug.cgi /webspace/bugzilla.example.com/show_bug.php
Alias /attachment.cgi /webspace/bugzilla.example.com/attachment.php

<Directory "/webspace/bugzilla.example.com">
    <IfModule mod_php5.c>
        php_flag engine on
    </IfModule>

    DirectoryIndex index.html

    <RequireAll>
        Require all granted
    </RequireAll>
</Directory>

Once you're happy with everything:

You can delete original-bugs-html...but it is safe to leave it. Please note that you won't be able to generate this data again unless you put the original Bugzilla install back in place, so it might be better to leave it here, in case you want to tweak the template or fix a script bug and rebuild the static pages, etc.

I have bugs that are just spam that bots posted.

Delete the static HTML files for them and the PHP code will report it as an invalid bug. Since re-running the script will regenerate the static HTML, you can also put a zero-byte file named BUGNUMBER-SPAM in the directory next to the static HTML (so bug number 6312 would have a file located at bugs-html/6/3/6312-SPAM). If this file exists, the PHP code will treat the bug report number as if it's invalid.

My bugzilla looked different.

You probably have to hand-edit template.html to match your site. This might take some effort. Pay attention for @symbol@ things in that file; those are replaced with bug-specific data when the static pages are generated.

What if this doesn't work?

This worked for my Bugzilla install, but the code is fragile in many ways, so a different version of Bugzilla, or some theme, or even just a quirk of a bug report that I didn't run into, etc, might break it.

If you have problems, file a bug, or better yet, send a patch, and we can talk it through.

Thanks!

--ryan.