Version 8 of BASH is available and fixes some bugs resulting in no data being flushed to BASH$HIST_ACTIVE_SESS_HISTORY. Again, thanks to Robert Ortel for reporting and tracking down the cause.
If you already have BASH installed, you can use the SQL*Plus update script update_v7tov8.sql.
- Fixed another UTC-conversion resulting (more...)
Version 7 of BASH is available and fixes two bugs leading to inconsistent rows in BASH$HIST_ACTIVE_SESS_HISTORY. Thanks to Robert Ortel for reporting and tracking down the cause.
If you already have BASH installed, you can use the SQL*Plus update script update_v6tov7.sql.
- Fixed a UTC-conversion bug around midnight, resulting (more...)
Version 6 of BASH is available. If you already have it installed, you can use the SQL*Plus update script update_v5tov6.sql.
- Made sure BASH works on Oracle 12c
- BASH is now compatible with RAC on Oracle 11.1 and higher. Each instance runs its own collector through a separate (more...)
Version 5 of BASH is available. If you already have it installed, you can use the SQL*Plus update script update_v4tov5.sql.
- Added a nightly purge job for historic data (defaults to 93 “days to keep”) (if you already created one yourself and use the update_v4tov5.sql update script you’ll (more...)
Version 184.108.40.2064 of Mumbai is available for download. Here’s the list of changes:
- Added ASH data viewer (supports ASH and BASH data)
- Added one-click 10046 tracing for scripts in the console window
- Lots of small tweaks and fixes
After releasing the BASH package that makes active session history data available without a Diagnostic Pack license, an ASH data viewer in Mumbai was the logical next step.
The layout as shown below is similar to what you might know from the ASH data shown in Enterprise Manager/Cloud Control/DB Console or from ASH Viewer, but there are, in (more...)
BASH – It’s ASH for the rest of us
Note: This is a historic blog post, that is no longer being updated. Please go to the BASH page for up-to-date information.
If you, like me, work with Standard Edition, Standard Edition One or don’t have a license for Oracle’s Diagnostic Pack, you probably miss Oracle’s ASH (Active Session History) badly. While Statspack, in my opinion, still can fill most of the gap the unavailable AWR leaves on SE/SE1, at least I miss ASH very much.
Like lots of DBAs and consultants, I am a huge fan of Tanel Poder’s snapper (more...)
Presenting at DOAG 2011 was a great experience, so I am very happy to present again at DOAG 2012. The title of the session is “Analysis and Visualizing Statspack and AWR data” (it will be in german language though).
Here’s the abstract:
Der Referent gibt eine Überblick über Statspack und Automatic Workload Repository (AWR), erläutert die Konzepte um die meist großen Mengen von vorhandenen Statspack/AWR Daten zu analysieren und führt in Demos vor, wie diese Konzepte in "Mumbai", einem frei-verfügbaren Windowsprogramm u.a. zur Oracle Performanceanalyse, implementiert wurden.
Der Referent zeigt die Probleme der Durchschnittsbildung, die Grenzen der Statspack/AWR Daten (more...)
No rocket science here, but since I use it so frequently, I thought this might be actually usefull for others, too.
Quote from the Readme.txt file:
WHAT IS THE POINT?
Once installed, the two tables, the PL/SQL package and the scheduler job
make it very easy to record regular snapshots of the results of SQL queries.
Since this is something I frequently need to do, e.g. for performance data
not snapped by STATSPACK/AWR, tracking a list of invalid objects in developer
schemas over longer periods of time, recording AUDIT information from remote
test/dev databases that are cloned from (more...)
Version 220.127.116.111 of Mumbai is available for download.
Here’s the list of changes:
- There is now a second page control on the left side of the main window where you can drag and drop pages to or you can right-click on a tab and select “Move to other side” to have a side-by-side view:
- The bind variables panel in the console window can now be collapsed:
- Package detail view: A view which combines the package and package body into one view with a procedure/function navigation tree and a list of errors in the package if there are (more...)
The databases I look after in my day job always showed a pretty low MREADTIM value, just a little above SREADTIM. Recently, the disks in our SAN were changed and we now have a huge caching module installed and now I consistently see MREADTIM values significant below SREADTIM. It seems that with MREADTIM ‹ SREADTIM Oracle reverts to calculating these values as it does when only NOWORKLOAD statististics are available. When doing this, Oracle ends up with an MREADTIM › SREADTIM.
I still don’t understand the rationale for this behaviour, since my idea of system and object statistics is that they should (more...)