After doing a full domain rebuild and Exchange 2003 Standard Edition migration for a customer, we found ourselves with a serious problem.

LSASS.exe and Store.exe during online maintainance would suddenly compete for CPU resources, and wouldn't stop.  Resulting in Store.exe needing to have a service restart every morning at 4AM PT after online maintanance was finished.  This was obviously a severe problem, with little to no known cause.

 For the migration, we followed Microsoft's recommendations of backing up the pub and priv dbs.  Installing Exchange with the /disasterrecovery switch and patching to identical levels, and then restoring.   Everything was flawless, all mailboxes intact, remapped users to mailboxes with no issues….but that was when this resource competition began!

So, we began testing, we used the migration wizard to move all the mailboxes to a new server, since we now trusted the domain and Exchange environment that had been rebuilt.  And suddenly, our resource competition disappeared!  Perhaps Microsoft hadn't actually tested a migration/recovery path quite as completely as they claimed?  So, now, I am testing the complete rebuild and recovery of the Exchange system to the same DNS name.