RE: High "log file sync" wait in HP-UX Benchmark ?
Date: Fri, 23 May 2008 06:58:40 -0400
Message-ID: <016501c8bcc3$f6f4c4a0$1100a8c0@rsiz.com>
Your average wait is only 9 ms. Your log file parallel write average wait is
only 2 ms (that is probably getting handled in raid 5 write cache, I hope
you're comfortable with that BAARF.)
I think you are referencing the earlier benchmark you did in January when the average waits were on the order of 10 times worse than that in average wait and much larger in number of log file sync waits. At that time you had an actual problem with the i/o configuration.
Take a look at JL's post from that time frame and analyze the effects of aggregation on your current situation. Very likely if you look at any individual transaction's session, it is not waiting significantly, but many of them are waiting a small amount in parallel. The way Oracle counts those waits in aggregation distorts the %Total Call Time in your report. JL's post referenced some convenient tools from Tanel and noted you don't have to do anything very sophisticated to see if you're looking at materially inflated numbers for waits.
Regards,
mwf
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org]
On Behalf Of VIVEK_SHARMA
Sent: Wednesday, May 21, 2008 4:00 PM
To: oracle-l_at_freelists.org; greg_at_structureddata.org; mwf_at_rsiz.com;
richa03_at_gmail.com; tim_at_evdbt.com
Subject: High "log file sync" wait in HP-UX Benchmark ?
Folks
During a Benchmark, High "log file sync" wait is observed for OLTP Transactions
Qs 1 What is the Solaris "maxphys" equivalent parameter on HP-UX 11.31? .. Increasing this had helped.
Qs 2 Will moving the Online Redo logfiles from Mounted Filesystem to RAW Devices help? ... This had significantly helped in the Solaris case since Oracle was using pwrite() function call to write to redo logfiles on Mounted Filesystem , while another function call (forgot which) was used to write to redo logfiles after they were moved to RAW Devices.
Qs 3 Since the Entire Set of Hard Disks is on RAID 5, could this be a significant cause for High logfile sync wait?
NOTE - Storage Box = HP EVA 600 SAN Storage Box
RAID 5 setup using SAME thereby consuming all of the 24 Hard Disks in the Box
Will provide Statspack Report as needed.
Cheers & Thanks in Advance
Vivek
P.S.
Config:-
Oracle 10.2.0.3
HP-UX 11.31
Application - Hybrid Banking Product
SP Report:-
Top 5 Timed Events Avg%Total
~~~~~~~~~~~~~~~~~~ waitCall
Event Waits Time (s) (ms)Time
- ------------ ----------- ------
log file sync 178,131 1,659 952.0
CPU time 97230.5
log file parallel write 140,197 240 27.5
cursor: pin S 888,008 147 04.6
db file sequential read 43,218 77 22.4
- CAUTION - Disclaimer *****************
This e-mail contains PRIVILEGED AND CONFIDENTIAL INFORMATION intended solely
for the use of the addressee(s). If you are not the intended recipient, please
notify the sender by e-mail and delete the original message. Further, you are not
to copy, disclose, or distribute this e-mail or its contents to any other person and
any such actions are unlawful. This e-mail may contain viruses. Infosys has taken
every reasonable precaution to minimize this risk, but is not liable for any damage
you may sustain as a result of any virus in this e-mail. You should carry out your
own virus checks before opening the e-mail or attachment. Infosys reserves the
right to monitor and review the content of all messages sent to or from this e-mail
address. Messages sent to or from this e-mail address may be stored on the
Infosys e-mail system.
***INFOSYS******** End of Disclaimer ********INFOSYS***
-- http://www.freelists.org/webpage/oracle-lReceived on Fri May 23 2008 - 05:58:40 CDT