Re: Statspack Interpretation

From: <mccmx_at_hotmail.com>
Date: 20 Apr 2006 05:51:57 -0700
Message-ID: <1145537517.413553.171480@i39g2000cwa.googlegroups.com>

Jonathan,

The figures don't seem to add up.....

2 of the top wait events are

                                         Waits      Wait Time (secs)
db file sequential read               330,557          2,156
db file scattered read                183,853            939

But the Elapsd Time (s) for the top SQLs ordered by 'physical reads' are no where near the wait times above:

Physical Reads Executions Reads per Exec %Total Time (s) Time (s) Hash Value
--------------- ------------ -------------- ------ -------- ---------

127,530 28 4,554.6 7.4 9.86 94.75 563927533

82,009 2 41,004.5 4.8 2.02 32.16 830612546

82,004 2 41,002.0 4.8 2.70 27.47 2625746475

80,298 2 40,149.0 4.7 2.50 35.60 914008173

78,425 2 39,212.5 4.5 4.14 60.78 386248711

The total time for the top 5 I/O SQLs is not even 10% of the total wait time for the i/o related waits...

Are you sure it isnt per execution times..?

Matt Received on Thu Apr 20 2006 - 07:51:57 CDT