-g ppf : isrd X bfrd

Hi , IDS 11.50 FC5W4 HPUX 11.11 On the onstat -g ppf output (+ awk script), sorting the data by isdel (checking what table/index have more deletes) I get this output: partnum lkrqs lkwts isrd iswrt isrwt isdel bfrd bfwrt seqsc rhitratio 0x01b00065# 160239022 0 1992544699 18424381 0 12145752 2299509993 31211360 0 100 0x01b00064# 540588991 0 2517874201 18424381 0 12145752 2765426072 31032567 0 100 0x01b00038# 147774601 0 0 18424381 0 12145752 286887491 34194569 0 100 On the third line, have the partnum 1b00038 (is an index) what have a lot of bfrd and 0 (zero) isrd . So, this isrd means read executed explicit by SELECT ? This bfrd have this value because of UPDATES/DELETES? In other words.. this index isn't useful for the system.. only take resources to keep it updated... and no SELECT use it ... (considering all statistics updated and OK). OR this isrd means the read access to disk/chunk , bringing the data to buffer pool ? what don't make sense for me, because the profile never has been reset (onstat -z) on this instance, since they start up Cesar

Used is from disk and bfrd is from cache. Don't know why zero used with no reset though. Art On Aug 18, 2010 2:45 PM, "Cesar Inacio Martins" < cesar_inacio_martins@yahoo.com.br> wrote: > Hi , > > IDS 11.50 FC5W4 > HPUX 11.11 > > On the onstat -g ppf output (+ awk script), sorting the data by isdel > (checking what table/index have more deletes) I get this output: > > partnum lkrqs lkwts isrd iswrt isrwt isdel bfrd bfwrt seqsc rhitratio > > 0x01b00065# 160239022 0 1992544699 18424381 0 12145752 2299509993 31211360 0 > 100 > > 0x01b00064# 540588991 0 2517874201 18424381 0 12145752 2765426072 31032567 0 > 100 > > 0x01b00038# 147774601 0 0 18424381 0 12145752 286887491 34194569 0 100 > > On the third line, have the partnum 1b00038 (is an index) what have a lot of > bfrd and 0 (zero) isrd . > So, this isrd means read executed explicit by SELECT ? > This bfrd have this value because of UPDATES/DELETES? > > In other words.. this index isn't useful for the system.. only take resources > to keep it updated... and no SELECT use it ... (considering all statistics > updated and OK). > > OR this isrd means the read access to disk/chunk , bringing the data to buffer > pool ? what don't make sense for me, because the profile never has been reset > (onstat -z) on this instance, since they start up > > Cesar > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > --001636eee445ffe204048e227000