Niall, both bugs referred to in this notes are fixed in 10.2.0.2 and they do not seem to lead to data loss or corruption (they resemble a lot what Alex refers to in his blog note about ASM). ...the point is whether you think those bugs were rooted in the architecture or where they simple code bugs. Do you really think a bug that forces all instances out during a rebalance was due to a code bug like an errant pointer? No. Those were implementation bugs. To a layman, the difference may seem moot if the remedy is the same (e.g., Opatch or patchset upgrade), but the fact is there are types of bugs. The bugs Niall pointed out are they types of bugs that get found when things go wrong. Unless your acceptance plan has a long list of such operations (e.g., RAC instance impact during offline disk rebalance) you are essentially judging the book by the cover. The thread that started this line of questioning came from me and I stated right up front that since ASM is optional software it is a matter of choice. It is important that forums like this look deeply into such options. It is the intellectually curious thing to do. -- //www.freelists.org/webpage/oracle-l