IIRC, there was a bug related to a shutdown abort during a kernel transaction that would cause corruption. I don't know the bug, but the additional info may help. One of the 'bugs' that was often discussed had to do with people switching between manual (rbs) and automatic undo after shutting down the database in an inconsistent mode (abort). Not a bug, but PEBCAK. Daniel "Adams, Matthew (GE Consumer & Industrial)" wrote: > > > Anybody remember to bug number for the problem > in early versions of Oracle9 where using undo > tablespaces could cause database corruptions. > (or is my memory playing tricks on me again) > > A brief search on metalink is not giving the > answer, but I'm still searching. > > Matt > > ---- > Matt Adams - GE Appliances - matt.adams@xxxxxxxxxxx > Nobody loves me but my mother, > and she could be jivin' too. > B. B. King