For some reason I thought that we stored instance level system stats since RAC nodes do not have to be identical in size/performance. The comment was a bit of thinking out loud, but let me research this a bit more. The mentioned sys.aux_stats$ contains predicate info, not system stats and is used by gather_*_stats to determine if a histogram should be considered. As Alex mentioned, the quickest route would be to see where the diff begins in the 10053 trace. On 6/6/07, fairlie rego <fairlie_r@xxxxxxxxx> wrote:
Thanks Greg, But doesn't gather_system_stats populate sys.aux_stats$ and hence is consistent throughout the cluster?
-- Regards, Greg Rahn http://structureddata.org -- //www.freelists.org/webpage/oracle-l