• It's like a thousand tools in your pocket •
The OSM is a complete set of Oracle database tuning tools based upon Time-Based Analysis and Unit of Work Time-Based Analysis. The OSM is SQLPlus based, non-graphic, and comes with absolutely no warranty. But, you will love it!
The OraPub System Monitor (OSM) is the tool kit referenced in our Video Learning seminars and Craig's book, Oracle Performance Firefighting. It is also used during our consulting engagements, technical research, and during our training classes.
To receive updates about OSM toolkit changes, sign-up for the Tools list.
Search Craig's blog for relevant posts at the link above.
There are a few unique tools bundled within the OSM tool kit. Here are the main ones:
- osm.sql is the menu. Really! Makes sense doesn't it?
- readmeEventHist.txt provides details about the collection, reporting, and charting of wait times.
- as.sql is a real time single session sampler. You repeatedly run the script (think: press /) and you can watch a session's activity.
- rss.sql is Craig's Real-Time Session Sampler. It is a simple but very powerful learning and analysis tool. Your selected active session details are written in real-time to a /tmp file, which allows you to stream the output by doing a "tail -f" on the file. On the command line you enter the filtering condition and the sample frequency. Craig uses this tool in his presentations, training events and also in his Video Learning seminar, Tuning Oracle Using An AWR Report.
- sqlelget.sql gathers SQL elapsed times so you don't get caught only using the average elapsed time. This is detailed in Craig's Video Learning seminar, Using Skewed Performance Data To Your Advantage.
- swhist[x].sql reports wait time details from v$event_histogram so you don't get caught only using the average wait time. This is detailed in our Video Learning seminar, Using Skewed Performance Data To Your Advantage. The data can be used in conjunction with Craig's Wait Event Time Distribution Analysis Tool.
- swgettimes.sql gathers individual wait times in real-time for a specific wait event so you don't get caught only using the average wait time. This is detailed in our Video Learning seminar, Using Skewed Performance Data To Your Advantage. On a super busy system you may need to use swhistx.sql instead.
- ttpctx.sql shows both wait time details along with CPU consumption merged into a single report giving you a better understanding of the time situation - key to doing an Oracle Time-Based Analysis.
- rtpctx.sql shows time with a response time focus. Like ttpctx.sql it shows total time, but it also shows time related to a single buffer get, physical IO read, commit, block change, etc. allowing you to do a more advanced Unit Of Work Time-Based Analysis (UOWTBA).
*** If a tool crashes, try adding the Oracle release to the end of the script. This is how we keep the older scripts working and available. For example, if rtsysx.sql crashes, try rtsysx11.sql then rtsysx10.sql or perhaps rtsysx8.sql.