On Thu, 2004-07-22 at 10:48, Trevor Tregoweth wrote: > after my last post, which i think wasn't quite to the point, i would like > to find out how to have a web-backup server, and how to configure them, so > that when one goes down the other takes affect.
I'm in the midst of deploying a high-availability system that does this using heartbeat software from http://linux-ha.org/ . The web service is handled by one machine only, with another one standing by as a hot spare. As James mentioned, the hard part is replicating data between machines. For static HTML, it's easy to set up a regular job synchronising the sites using rsync. For databases it's a little tricker. I'm experimenting with using pg_dump and pg_restore to periodically move data between machines. It's clunky, and losing a small amount of data in event of a failure is likely, but it falls well within my requirements. Your mileage, of course, may vary. -- Pete -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
