R
Richard Banville
Guest
Yes, disk I/O is an expensive operation. Without going through an entire tuning exercise here… Are the physical writes to the files really the bottle neck or just a frequent operation? Are you running with page writers (apw/biw/aiw)? They should be taking care of most of the write I/O for you. Are these areas being frequently extended – more space requested at runtime? Are your checkpoints in check? Is your physical database layout configured properly for such growth? (area configuration, file location, RPB, toss limits etc) Are your indexes 100% compacted which results in expensive index block splits? What does the “blocked clients” screen of promon tell you? From: bremmeyr [mailto:bounce-bremmeyr@community.progress.com] Sent: Monday, May 04, 2015 9:42 AM To: TU.OE.RDBMS@community.progress.com Subject: [Technical Users - OE RDBMS] Controls to help high demand for record create Controls to help high demand for record create Thread created by bremmeyr What measure can be used to show if the database is hitting a bottle neck during periods of high create record activity? During this event all users are impacted with slow database response time. What I am trying to get to is visibility to the bottle neck. Is the issue at the DBMS or hardware level? Trying to identify what DBMS adjustments can be made to help minimize the impact on other users during these periods of high load? With this case, resource Monitor is showing the most active disk I/O write session is prowin32.exe to 3 extents of the database. 1 is the index extent and the other is a data extent. OpenEdge Enterprise DB, 10.2B, Windows 2008 R2 Stop receiving emails on this subject. Flag this post as spam/abuse.
Continue reading...
Continue reading...