We use BusinessObjects to query an Exadata back end. Recently, we had an issue where the overnight DW load failed entirely - this was explained as a "resource locking issue caused by a long-running BusinessObjects report".
Does this sound plausible to the people here? My understanding is that both BusinessObjects and Exadata have preventative measures available to prevent such issues arising.
Please mind there is way too few information to truly know what happened. It requires at least traces of both operations to understand what truly happened.
One thing which could have happened is the following:
- if the BO reports use smartscans/offloading, they do full scans on the segments in the query via direct path.
- data loading requires access to the segments it's about to load or change. This can not be done when these segments are scanned via direct path.
Regular (non-direct path) access does not require the same protection from the database, so loads can co-exists with them much better.
If you can get more information about where the load did wait for, we can do a little more analysis.