Re: Execute Count in AWR Report
Date: Sat, 18 Jul 2020 19:05:06 +0100
Message-ID: <CAGtsp8=6pO8NY5nYNA3t_FXSr_Kwa+9cMT7hS+F6HBrqgibJQA_at_mail.gmail.com>
My first question is why isn't the first run counted in the AWR?
The AWR takes a snapshot of what's currently in the library cache, the report then calculates the difference between one snapshot and the next. In your case a viable scenario would be:
12:00:00 - statement not in the library cache 12:04:51 - statement executes for the first time and gets into library cache 12:07:31 - statement finishes first execution and is no longer pinned inlibrary cache
12:10:00 - statement is flushed from library cache 12:13:21 - statement executes for the second time and is reloaded into the library cache
12:14:52 - statement finishes first execution and is no longer pinned in library cache
13:00:00 - statement is still in the library cache and is captured in the snapshot
AWR reports difference between snapshots - the first execution was never captured, the report shows only stats for the second execution.
My second question is whether the other reported valuers for the statement
such as User I/O wait time , buffer gets, and disk reads are also only for
the second run?
I would expect so.
You could run of the awrsqrpt report for the statement for the interval and
see what that said about number of executions, work done per execution etc.
Regards
Jonathan Lewis
-- http://www.freelists.org/webpage/oracle-lReceived on Sat Jul 18 2020 - 20:05:06 CEST