> See below for a 'select count(*) from dba_extents' on an Apps database > (total of 217,167) extents sitting on 452 datafiles, and the snapshot of > V$SESSTAT and V$SESSION_EVENT for that SID. The query too about 13 minutes > (dev server, slightly slow disk). Interesting stats to note? Look at > 'revursive calls', 'CPU used%', 'cluster key scans' and 'cluster key block > gets' stats, 'session logical reads', number of dbfile reads. Most were > against the C_FILE#_BLOCK#, UET$, SEG$, SYS_IOT_TOP_132603, I_FILE#_BLOCK#, > OBJ$, C_TS#, TS$, FET$ objects among others. Which version of db? Was it fully LMT or only partially? In full LMT configuration you shouldn't have any records in UET$ and FET$ (which contents can be cached in dictionary cache), instead there is IO on x$ktfbue (used extents), which in turn goes to segment headers... But yeah, select count(*) from dba_extents, especially in Apps like environment kill performance and yet there are monitoring tools around which query it after every five minutes for some reason... Tanel. ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx put 'unsubscribe' in the subject line. -- Archives are at //www.freelists.org/archives/oracle-l/ FAQ is at //www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------