ProTop 3.3t is on the loose!


.. and is available for download at ProTop

Release notes are in etc/NOTES_3.3t but highlights include:

Fixed an undocumented feature whereby the dashboard summarization of resource and latch waits was improperly initialized on every sample.

Added "as of" to the ProTop version so that multiple builds with the same version id can be distinguished (this is for development builds).

Modified the "*" screen to no longer "toggle" Client Statement Cache but rather explicitly turn it ON, OFF or not change it (?). Also made the warning text RED (see above).

Fixed a couple of "quirks" in the profiling code (the "y" command) and made the warning text RED.

Fiddled with bin/ Improved the formatting of the timing and suppressed messages if being run from cron so that it can be run automatically and log the results more usefully.

Added syncIO to the dashboard - this is NOT updated continuously like ioResponse. But it will reflect the most recent value that has been logged in log/syncio.log so if you run bin/ nightly (or more) the most recent value will be shown.

Added index overlap report to the ^D/^R dictionary report. This report looks for indexes with common components in order to indentify redundant indexes that could potentially be eliminated. Two output files are created in the report directory:

friendlyName.idxoverlap.rpt # analysis showing overlapping components
friendlyName.redundant.rpt # completely redundant indexes

(This actually predates the "Communities" thread -- really!)

Fixed a bug where the new automatic .pf file detection in dbmonitor.p was being a bit over-exhuberent and adding extra pf file references for each additional resource being started. This doesn't hurt anything but it does look funny.

Modified various fields on the "config" screen to reflect online parameter changes. There is more work to be done here but a lot of the big ones (-spin, -B...) are covered now.

Added the "C" command to clear Client Statement Cache settings that this ProTop session has made. Users will be reminded on exit if uncleared sessions remain.

Enriched the "dirty schema" alert to list the first 10 offending objects.

Added a warning at startup if no dbanalys can be found. Included instructions for remedying the problem.

Added hashPct metric to track the value of -hash as a percentage of the sum of -B and -B2. Added an alert if that percentage is less than 25%. (OEM has some bad UI behaviors.)

Adjusted -pica to reflect the command line value. PROMON and the .lg file report MB, _dbParams reports bytes. The startup parameter is specified in KB.

Added "L": Login Broker Config. This reports on 4gl and SQL broker configuration and helps you to ensure that you have properly set -Mn, -Mpb, -Ma, -SQL* etc. (requires VSTs for OE11.5+, so make sure you run "proutil dbName -C updatevsts"!)

Added the many various operating system environment variables that control behaviors to bin/localenv.x and documented their usage. To use or modify these features copy the appropriate variable into bin/localenv (no ".x") and set accordingly. bin/localenv.x may be updated with new releases so do not use it directly - that's why it has a ".x". The proper script for localizations is /bin/localenv - the installer will not overwrite that.

Added "&", "ProTop Properties" screen to show the current values of many of the variables described above. However, not all of those variables are used by protop.p or pt3agent.p and some are not globally available. Some of the variables can even be safely changed "on the fly" if you'd like to experiment.

Updated the help screen ("?"), reassigned some of the more obscure command keys and added some previously "lightly documented" commands.

Added an "ai info" screen for a detailed look at ai extent status. Use the control-a key to bring up this screen.

Changed the default for PCTLASTX from yes to no. Basically nobody actually monitors full extents by default. If you really want to do that then just set it to "yes" in bin/localenv.

Added some interesting columns to the app server data (first configure etc/appsrv.cfg and then use the the "@" command). MaxSrv is the maximum number of servers configured in, Srvr% is the percentage of that maximum that you are using. Likewise MaxCli is the maximum number of client connections and Cli% the percentage of that maximum that you are at. An ActCli column has also been added to show the number of active clients.

Shuffled around the functions of ^d and ^r -- ^d now generates dump & load scripts. ^r generates reports.