onstat g opn - lockmode

Hi , Ifx 11.50 FC9X6 , AIX 6.1 (with 1 RSS read-only synchronized) Monitoring the instance , I'm trying figure out why one stock historical table persists keep 60% of the table size into buffer pool a few days (checking with onstat -P) (60% = 10GB). So, I'm looking some historical "snapshots" what I create each 15 minutes (onstat commands...) for users what is reading constantly this table or executing sequential scans. for my surprise with "onstat -g ppf" I detect the bufwrites growing for this table. This table should not have any update at this epoch of month.... is a historical stock, they was populate only between months. At the same time, the "onstat -g opn" show me just one user accessing it. Tracing the session , this user was running a report application, so , no update/insert/delete into the code (4GL). Looking the last 15 minutes (the "snapshot" what I capture) , only the same user was accessing this table. onstat -g ppflog.2012-07-21_1400.gz : partnum lkrqs lkwts dlks touts isrd iswrt isrwt isdel bfrd bfwrt seqsc rhitratio log.2012-07-21_1400.gz : 0x2700004 52183079 0 0 0 439858377 9530760 0 10665012 497228308 27129013 4 97 log.2012-07-21_1415.gz : 0x2700004 52183079 0 0 0 >> 442757465 9530760 0 10665012 500191255 27129533 4 97 log.2012-07-21_1430.gz : 0x2700004 52183079 0 0 0 446671162 9530760 0 10665012 504198331 27130293 4 97 onstat -g opn log.2012-07-21_1400.gz : tid rstcb isfd op_modeop_flags partnum ucount ocount lockmode log.2012-07-21_1400.gz : 293488390 0x0700001d1527a108 17 0x00000400 0x00000007 0x02700004 1 1 7 log.2012-07-21_1415.gz : 293488390 0x0700001d1527a108 17 0x00000400 0x00000007 0x02700004 1 1 9 log.2012-07-21_1430.gz : 293488390 0x0700001d1527a108 17 0x00000400 0x00000007 0x02700004 1 1 9 I'm trying figure out : 1) Why the lockmode show the status 7 and 9 ( 7 Intent exclusive lock , 9 Exclusive lock) if have only selects over the table. 2) Why op_mode have flag 0x400? (0x000400 Manual record lock). 3) who is updating this table!??? ( could be the btree scanner ? And we can't able to detect here they accessing the table and updating the rowids? The application/report (4gl) what the user is running : - There is no update statistics into the code - no lock command / transactions - the session use only dirty read isolation - any "set environment" defined (like batched read) any tips? My primary objective is discovery why this table don't leave the buffer pool and for that , discovery who is "updating" this table... Regards Cesar