On the restoration and updating of a Drupal 4.7 based website recovered from archived backups.

I didn't keep great notes as I worked through this project and I did pretty minimal research - I was inspired to give it a go one day and I just went for it. So I'm doing my best to put a coherent work log together from memory.

Page is under construction, all content subject to change.

introduction

This site has gone through a number of iterations over the last couple of decades, with a long stretch of being offline in the 2010's and early 2020's. In 2004 I moved the site from Greymatter to Drupal. I wanted to make a "community" site where I could post and share content with friends and family and they could do likewise. I maintained the site for a few years in this guise and abandoned it sometime after 2008. The site files and database were backed up and copied to a CD (of course) and left to rest.

As I've been working on the latest version of the site, I've been thinking about those previous incarnations. Recently, out of curiosity and nostalgia, I decided to browse the Internet Archive Wayback Machine to get a feel for the timeline of the ultrasciencelabs.com domain. And while browsing I realized how much I left abandoned. And then then the thought struck me- "I've got backups, would it be possible to restore the site, or extract the content?" - or maybe even update/migrate it to a current Drupal install.

I dug though my archives and found the last full backup of the site dated from 2008 before it went dark. I had no idea what functional state it would be in, but I knew to have any chance of working it would likely need to be restored to an environment of the time.

research

I figured the most convenient way to test the backups would be to fire up a virtual machine with an Ubuntu install and load up a LAMP stack configured as "localhost". Transferring files between my host machine and the VM would be much faster than trying to upload/download via FTP, and by keeping everything on a local VM it would be faster/simpler to add/drop databases, modify file permissions, or even wipe out and rebuild new VMs as needed as the site version updates progressed.

The CHANGELOG.TXT file in the backups gave the Drupal version as 4.7.4 which was released in October 2006. Per the INSTALL.TXT details it would need PHP version 4.3.5 or greater or PHP 5. Per Wikipedia's Ubuntu Version History 6.0.6 LTS was released around that time and came with php 5 packaged. But since my goal is updating the site I figured it might be good to use an Ubuntu version with a more up-to date php version that would support newer versions of Drupal and 6.06 was pretty early in the Ubuntu run, I wasn't sure how nicely it would play with VirtualBox. I rolled the dice and went with 8.04 LTS.

implementation

Ubuntu 8.04 installed fine, and running apt installs for PHP, MySQL, and Apache were easy enough. But I was having a tough time with the VirtualBox Guest Additions - specifically the shared folder setup. Easily moving data on and off the VM was mission critical, but I couldn't get the extensions working - re-installs, configuration troubleshooting, nothing was working. I suspected the issue was likely due to the packaged version of the extensions being so out of date.

So I decided to migrate to a new Ubuntu version. But which to choose? Looking at the Drupal releases version 6.28 was released in 2016 and versions 5.0 thru 6.28 all listed php 5 in the requirements. So I went with 14.04 LTS figuring it might be new enough to play nice with VirtualBox, but not be too new for Drupal 4.7.x, but also be new enough to support Drupal 5.0 and up.

So I rolled a new VM, installed my LAMP stack as before, and this time I was able to get the VirtualBox Guest Additions working. Instead of using the packaged version, I downloaded the most recent version that listed Ubuntu 14.04 in the supported installs - 6.1.18. That did the trick and I was able to get a VirtualBox shared folder working, and now the real restoration work could start.