I think I’ve finally gotten the blog formatting to be like I want … The syntax highlighter plugins were not good enough as I want very flexible output formatting (not just code snippets, but sqlplus script output) as below:
A new version 2.0 of the XPLAN_ASH utility introduced here is available for download.You can download the latest version here.The change log tracks the following changes:- Access check- Conditional compilation for different database versions- Additional activity summary- Concurrent activity information (what is/was going on at the same time)- Experimental stuff: Additional I/O summary- More pretty printing- Experimental stuff: I/O added to Average Active Session Graph (renamed to Activity Timeline)- Top Execution Plan Lines and Top Activities added to Activity Timeline- Activity Timeline is now also shown for serial execution when TIMELINE option is specified- From 18.104.22.168 on: We get the ACTUAL DOP from the undocumented PX_FLAGS colu
Jagjeet Singh mentioned that he has built a V2 of his cool SQL Dashboard sqlplus-based monitoring script, which now also shows a nice colored ASH graph output. Yes, right inside sqlplus!
I’m happy to announce Enkitec TV which is a video channel/aggregator of some of the coolest stuff we do here at Enkitec. I have uploaded some of my videos there, including the previously unpublished Oracle parameters infrastructure hacking session and Kerry’s & Cary’s E4 Exadata interview is available there as well!
Note: This blog post actually serves three purposes:
Ok, I’ve wanted to write this blog entry for a long time – and now it’s time!
Most of my blog readers (thank you!) are performance-minded computer enthusiasts, who care about efficiency and optimization. You’ve been tuning SQL execution plans, instance and OS configuration so that your sessions would achieve the same results with less work and also with less waiting!
You probably know to appreciate why asynchronous I/O must be enabled for busy modern databases, so that your database sessions can do I/O (talk to the storage) without actually having to wait for the I/O operations to complete! You can increase the processing throughput, by not submitting every single I/O separately and waiting for it to complete, before being able to process the results and submit the next one. Asynchronous I/O is a crucial thing for good performance.
So this will be my Oracle related Christmas present for you: A prototype implementation that extends the DBMS_XPLAN.DISPLAY_CURSOR output making it hopefully more meaningful and easier to interpret. It is a simple standalone SQL*Plus script with the main functionality performed by a single SQL query. I've demoed this also during my recent "optimizer hacking sessions".
DBMS_XPLAN.DISPLAY_CURSOR together with the Rowsource Statistics feature (enabled via SQL_TRACE, GATHER_PLAN_STATISTICS hint, STATISTICS_LEVEL set to ALL or controlled via the corresponding hidden parameters "_rowsource_execution_statistics" and "_rowsource_statistics_sampfreq") allows since Oracle 10g a sophisticated analysis of the work performed by a single SQL statement.
This is not really anything new - in fact Tanel Poder has already blogged about it a while ago. Tanel has specifically covered the handling of "urgent" TCP packets and how this could be used to signal a "cancel" to another process, however this only works on Unix environments and not with Windows SQL*Plus clients. In Tanel's article it is also mentioned that there is an officially documented way of doing this via the Resource Manager if you happen to have an Enterprise Edition license.
In my quick tests however the call to DBMS_RESOURCE_MANAGER.SWITCH_CONSUMER_GROUP_FOR_SESS using "CANCEL_SQL" as consumer group only errors out with ORA-29366 saying that the specified consumer group is invalid.
If you work with SQL Trace files (and profile them) then you should check out the awesome novel use of the “external table preprocessor” feature explained by Adrian Billington here:
Ironically just a day after writing my “Evil things” article, I noticed a note in MOS about how to enable an event 10384 at level 16384 to get a parallel plan to be executed in serial:
Relax, I’m talking about the Oracle Database kernel here, not the corporation ;-)
Here’s a couple of more reasons why not to play around with undocumented debug events unless you’re really sure why and how would they help to solve your specific problem (and you’ve gotten a blessing in some form from Oracle support too):