Hi Greg One change is in the way stats are collected. This is Oracle Apps 12.1.2 and we use fnd_stats.gather_schema_stats. In solaris it was, SCHEMA_NAME => ALL Estimate Percent => 10 Degree => Null Backup Flag => NoBackup Restart Request ID => Null History Mode => Last run Gather Options => Gather Modifications Threshold => Null Invalidate Dependent Cursors => Y in Linux it is CHEMA_NAME => ALL Estimate Percent => 10 Degree => Null Backup Flag => NoBackup Restart Request ID => Null History Mode => Last run Gather Options => Gather Auto Modifications Threshold => 10 Invalidate Dependent Cursors => Y But my understanding was this is better because it only collects stats for any table if more than 10% of the data changes and this reduced stats gathering from 9 hrs to 30 minutes. To minimize issues with version differences we maintained the version of rdbms s (11.1.0.7) between solaris and linux and another change to init.ora was not to use memory_target and memory_max-target (that should not matter I hope in this particular case thoug h. We are using sga_target and sga_max_size ) Thank you Kumar On Mon, Jun20, 2011 at 5:31 PM, Greg Rahn <greg@xxxxxxxxxxxxxxxxxx> wrote > 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 >