This sounds like a bug with a hash group by. I would recommend working with Oracle Support. Bug/patch 6471770 would be where I would start. On Wed, Aug 26, 2009 at 9:20 PM, Taral Desai<taral.desai@xxxxxxxxx> wrote: > Hi Gurus, > We are using oracle 10.2.0.3 on sparc solaris 64 bit with 16 CPU and 32GB of > physical memory. Our sga_target is 12G and pga target is 15G. > We are receiving > ORA-32690: Hash Table Infrastructure ran out of memory > We had identified query which is running in parallel 8 and using huge hash > group by. May be due to that this error is there. But here are some > questions > 1. How to find why this error is coming. Any trace dump or pga dump can > help ? Because i believe it was not using temp space at all. We have 2 TB > temp space free. > 2. i think for hash group by if it don't get any memory it spill to temp ? > then why this error > 3. How to monitor this while this error is happening. > 4. How to dig into issue using dumps and what to look for. > 5 IS PGA_AGGREGATE_TARGET only use 5 % for each parallel process -- Regards, Greg Rahn http://structureddata.org -- //www.freelists.org/webpage/oracle-l