Rick Hurst Web Developer in Bristol, UK


plone migration woes

At the moment, my pet plone project DFR skate zine is still live, but because netsight have some devious plans for the server it is on, I am trying to upgrade it to the latest version of plone. The site is current running plone 2.1.1 (tip: if you can’t remember what version of plone your site is running, go into the ZMI and click on portal_migration – you’ll find lots of useful info there, including the plone version).

Initial attempts didn’t go well, but I haven’t given up yet. I’m now trying a different approach – creating a blank plone 2.5 site and trying to bring the content over bit by bit. Some of it imported just fine and other objects are throwing various key and attribute errors.

This blog post by Andreas Jung looks like it could be quite useful – my version jump is nowhere near as big as the version described here, but the technique looks interesting – particularly the script for cloning a tree of objects.

I have some fairly radical plans for the development of the DFR site, I was actually thinking of recreating the skin and content type products from scratch, because a) I want a single skin for public and admin view now and b) because the archetypes content types were patched together in a bit of an ad-hoc way, and now they seem a bit.. well… unclean, and basically I think I can do better now..

so… the masterplan at the moment is to extract the content into some format that I can get to the data easily (as I still find it very confusing trying to get to data “lost” in zodb, when the site is broken), then write a script to populate the new site from the data. Obviously xml fits the bill (along with files and folders for media content).

Marshall and XMLForest have been whispered to me on a need to know basis! If I can crack this i’ll be much more confident about maintaining plone sites long term, in the same way that I am confident with CMS’s that use a SQL database for storage.