The temp usage changing is more certainly from plans changing - nothing to do with the OS directly, so it would make more sense to focus on why that is -- stats, db version, parameters, etc. On Mon, Jun 20, 2011 at 4:17 PM, David Fitzjarrell <oratune@xxxxxxxxx>wrote: > Are your SGA/memory settings the same on Linux as they were on Solaris? > To what values are sort_area_size and hash_area_size set? Have you > monitored who's using the TEMP space, how much they're using and what > statements are the largest offenders? Is the TEMP tablespace sized the same > on Linux as it was on Solaris? Have you traced sessions with strace? > > *From:* Kumar Madduri <ksmadduri@xxxxxxxxx> > *To:* oracle Freelists <Oracle-L@xxxxxxxxxxxxx> > *Sent:* Monday, June 20, 2011 2:26 PM > *Subject:* Temp usage on Linux vs solaris > > Hi > We recently migrated our ebiz solaris db to linux. We have some bad queries > in solaris which take up temp segment (I mean mor than what it should take > if the query was properly written in the first place). > Since we migrated to LInux, the queries are as bad as ever but we see more > ORA-1652 errors now than we were on Solaris. > We are trying to identify the sql_id and report that to developers. But > have any of you noticed increased temp space usage after migration to linux > from solaris. > > Thank you > Kumar > > > -- Regards, Greg Rahn http://structureddata.org