In big projects (and in small ones, too) strange things happen. Stuff breaks in ways that you can never ever imagine before. Failure.
If failure happens you have to deal with it. Fix it. Go on. That's what your job is. Failure or disaster is not that fatal is it might seem at the beginning [2].
Failure - by the way - is normal if you are creative. If you want to get things done and moving you will fail. The only... wait... THE ONLY thing that should NEVER EVER HAPPEN IS: the same failure happening twice.
Learn from your failure. Don't repeat it.
Google openly uses the notion to publish the reasons for failures in their "post-mortem" reports. For instance 2010 their App Engine Service went down. The engineers at the sites had to cope with the (huge) problems, solved them after a while and published a long "post mortem" report [1].
That's a great way to build trust in your organization. Plus you are owning your own bad news.
Update
- July 2011: Another nice post mortem report from Google at [3].
[1] https://groups.google.com/group/google-appengine/browse_thread/thread/a7640a2743922dcf
[2] http://sethgodin.typepad.com/seths_blog/2011/06/disaster-tolerance.html
[3] http://googleappengine.blogspot.com/2011/07/postmortem-java-app-engine-outage-july.html