Running backups of wordpress.com blogs
Samstag, 16. Juni 2007, 21:22
I’ve been confronted with the task of securing a wordpress.com blog so that any inconvenience resulting in loss of data can be avoided, be it through a malfunction in the wordpress.com software or through hackers. Also, if wordpress.com one day decided to censor specific topics, one should be able to move fast and reestablish the whole blog on a different platform.
The first part to securing the blog is provided by wordpress.com itself: An export-tool can be found under
"Manage" > "Export"
There one can download an extended rss-feed file that contains not only all posts, but also categories, authors and meta-data of attachments. The only thing missing is the blogroll, configured plugins and of course the theme of the blog. The blogroll should be included in the export in future versions, since this is something that can change rather rapidly, but the other two would be a “nice extra”.
Either way, each time you want to backup the blog-data, you’d have to login, click on manage, then export, then download the file.. Nachtwächter (German) has published a script with which one can automate this process and which can be scheduled using cron. Unfortunately the script doesn’t import the images and other attachments that are being used on the blog.
In comes the Improved Wordpress to Wordpress Importer developed by zappoman. This importer will download all the attachments off of the old server and include them as attachments into the new blog.
But alas zappos’ script seems to be designed for an online migration to a new domain. That’s why it remaps the old links inside the posts, which is quite nifty, but isn’t really necessary if one wants to simply redeploy the site using the same domain-name or simply build a backup-system.
So, i guess i’ll be making some changes to the script, that Nachtwächter created. Those changes would have to include an xml-parser so that one can determine which attachments should be downloaded and simply put the files in a directory. And if anybody has an idea, how to download the list of links, then please share!
Tell me more about what you want to do with my plugin, I’m sure it could be modified pretty easily. My script is basically inspired by the Wordpress to Wordpress importer that comes with standalone Wordpress and WordpressMU, and so it really as you said is designed to import a wordpress XML file (from the export you mention) into a wordpress install. Of course it’s written under creative commons licensing and so the Wordpress.com team could use it on wordpress.com but someone would need to convince them to do so.
Anyway, I think the combination of the Nacthwachter tool and my tool would be a great solution for keeping a backup ready for redeployment.
@zappoman: well basically i’d like to a have a backup-system, that stores everything. This means that the script of Nachtwächter would have to parse the xml-file and then download all attachments (or only the new ones) and store them in a folder.
If/when the original server crashes, the import script wouldn’t have to download the files, but it would ask where they are (on the server, if you have direct access) or give you the opportunity to upload the files one by one. It would then have to put them in the right place, recreate the thumbnails and link them to their respective database-items.
The first step would be to include your algorithm of downloading the files into Nachtwächters’. After that the task of reimporting those files is a little bit easier (at least when the user can access the server directly, i.e. via ftp).
“Also, if wordpress.com one day decided to censor specific topics, one should be able to move fast and reestablish the whole blog on a different platform.”
Wordpress can do that, although you have your blog on your own domain and server??? How would they do that?
I have been using serendipity, but I am about to recommend Wordpress to colleagues of mine.
No, actually Joerg, I’m talking about wordpress.com which is the hosting site for many bloggers. Of course, when you have your own wordpress-installation on your server, wordpress.com can’t censor that!
