Anyone have any info on what exactly "Disk file operations I/O" is? A customer upgraded from 10.2.0.4 where things were running fine to 11.2.0.3 and now they are hitting high contention on buffer busy waits. The buffer busy waiters are waiting for a session that is waiting for "Disk file operations I/O". Here is an example of 4 minute period where 5 sessions backed up on buffer busy waits. Both waiters and blocker are executing the same select for update. All users SQL_EXEC_ID stays the same so it's the same execution (ie not multiple executions) Buffer Busy Waiters: MAXST MINST COUNT(*) SESSION_ID P1 P2 SQL_ID BSID ----- ----- ---------- ---------- ---------- ---------- ------------- ------- 09:54 09:58 26 1526 1812 278732 3gbsbw6w8jdb3 11 09:54 09:58 25 528 1812 278732 3gbsbw6w8jdb3 11 09:54 09:58 25 1514 1812 278732 3gbsbw6w8jdb3 11 09:54 09:58 26 777 1812 278732 3gbsbw6w8jdb3 11 09:54 09:58 25 33 1812 278732 3gbsbw6w8jdb3 11 Blocker, sid 11: MINST MAXST EVENT COUNT(*) SQL_ID BSID ----- ----- ------------------------------ ---------- ------------- ------- 09:54 09:58 Disk file operations I/O 26 3gbsbw6w8jdb3 The wait is on a data block of a table. The waiters buffer busy wait data looks like (all same P1 and P2) ST P1 P2 CURRENT_OBJ# FN BLOCKN SQL_ID BLOCK_TYPE BSID ----- ---------- ---------- ------------ ------ ---------- ------------- --------------- ------- 09:54 1812 278732 34395 1812 278732 3gbsbw6w8jdb3 data block 11 34395 is a table, applsys.fnd_concurrent_request (db_files is 10000) The query in question is like: SELECT blah blah FROM fnd_concurrent_requests R, fnd_concurrent_programs P, fnd_application A, fnd_user U, fnd_oracle_userid O, fnd_conflicts_domain C, fnd_concurrent_queues Q, fnd_application A2, fnd_executables E, fnd_conc_request_arguments X WHERE blah blah FOR UPDATE OF R.status_code NoWait / - Kyle Hailey -- //www.freelists.org/webpage/oracle-l