My Intralink 8 server went down last year and lost everything. Fortunately I do cold backups every night of Oracle, LDAP and the vaults. I also export the LDAP tree structure periodically. It took a couple of days of messing with it, but I got everything back and working. When you do your rehost make sure your target Windchill installation is the same build has your source installation. It's a good idea to test your backups a couple of times a year, if not more.
Following are some options, there can be many more to consider:
a) Keep the Windchill Site name separate from the actual host file name. For example,
b) Use a HTTP proxy server for Windchill, this will allow you to effortlessly switch Windchill hardware in the event of a failure. HTTP Proxy is easy to setup.
c) Maintain a test environment which is a replica of your production environment. Periodically rebuild your test server from production
d) If your production server is a monolithic instance, it is advisable to have a HTTP proxy. This will allow a quick redirect to the test server (re-hosted from production)in the event of a catastrophic production hardware failure
e) If you have a separate Application server and Database server (dedicated to Windchill only), you might want to consider installing all Windchill components on the database server to make it a monolithic instance. This will add redundancy to your existing production hardware platform.
e) If your production server is a cluster, obviously you have built application redundancy, I still recommend a test environment with periodic re-host from production