WordPress WXR import - importing a large file

Question

I am in the process of working on a project involving the migration of old custom driven cms data into a WordPress powered site.

To do this migration, we have decided to convert the data from the current site into a WXR import file.

The issue we're facing though is that the resulting WXR file is about 100mb in size. (about 10000 posts, of varying types and about 80000 attachments)

The import is obviously a very slow process given a file that size, so my question is:

Does anybody have any experience of importing a file that large?

If so, what can be done to speed up the process? More memory, more cpu? etc?

My current idea is to spin up a large EC2 instance and try the import on there, and then export the resulting MySQL to the final hosting location.

Answer 1

with 80k attachments you will waste time mostly on transferring the attachments, the network delay will be your killer. If you can avoid importing attachments at all that will probably be the best, and after the WordPress is setup you can try to find a plugin that scans a directory and creates attachments for its file (sorry forgot the name of it). Otherwise set your staging wordpress in the same LAN as the source sever.

In any case I remember reading that amazon AWS services have relatively slow disk access times, so you might better get a powerful VPS from companies that bill by the hour. I think rackspace do that.