Hi, All,=20 We have a database with buffer busy wait that we can't explain.=20 A huge table (1.5TB) with long raw column and high concurrency insert. (6-9M rows inserted per day) No FK etc. Just one PK and two non-unique indexes.(created_date and end_date column) Table/index has set the freelists=3D23. But we are seeing high buffer busy wait event for the insert session. The reason code is mainly 120(concurrent read) and 220(concurrent modify) What is the root cause for this buffer busy wait? Could it because index problem? Can someone explain? Is there any solution to this situlation? SQL: =20 insert into gallery ( item_id, x_size, y_size, start_date, end_date, thumbnail, thumbnail_len ) values ( :item_id, :x_size, :y_size, TO_DATE(:start_date, :"SYS_B_0"), TO_DATE(:end_date, :"SYS_B_1"), :thumbnail, :thumbnail_len ) call count cpu elapsed disk query current = rows ------- ------ -------- ---------- ---------- ---------- ---------- -----= ----- Parse 0 0.00 0.00 0 0 0 = 0 Execute 837 11.24 185.55 1130 91 35956 = 825 Fetch 0 0.00 0.00 0 0 0 = 0 ------- ------ -------- ---------- ---------- ---------- ---------- -----= ----- total 837 11.24 185.55 1130 91 35956 = 825 =20 Misses in library cache during parse: 0 Misses in library cache during execute: 1 Optimizer goal: CHOOSE Parsing user id: 18 =20 Elapsed times include waiting on following events: Event waited on Times Max. Wait Total Wait= ed ---------------------------------------- Waited ---------- ----------= -- SQL*Net message from client 837 5.14 163.= 56 buffer busy waits 31720 0.12 158.= 55 db file sequential read 1125 0.10 12.= 39 latch free 474 0.03 3.= 33 log file sync 918 0.76 3.= 08 SQL*Net message to client 836 0.01 0.= 01 SQL*Net break/reset to client 24 0.01 0.= 01 db01$>grep "buffer busy waits" prod_ora_10686.trc|grep "p3=3D120" |wc -l 27536 db01$>grep "buffer busy waits" prod_ora_10686.trc|grep "p3=3D220" |wc -l 4183 --=20 Regards Zhu Chao www.cnoug.org -- //www.freelists.org/webpage/oracle-l