Disclaimer
The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.
Liability Disclaimer
In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.
Posting
Within a single OSPF domain, without any redistribution with other routing protocols, the only two situations I can think of where you might see a routing loop if is OSPF is "broke" (e.g. some kind of bug) or perhaps during OSPF convergence (when not every router's topology database agrees). The latter, though, is often very brief. Not much you can do to totally preclude loops during convergence, but on Cisco platforms you can tune some parameters that will support faster convergence.
Regarding dealing with broken OSPF, avoiding bugs is sometimes aided by avoiding special code releases and/or new "new" feature code releases. Also, Cisco devices generally don't deal well with low memory, so you want to insure you have adequate memory.
