topic Re: Oracle Logs in Disk Enclosures

Oracle Logs

Thomas Maddox — Thu, 17 Apr 2003 12:36:22 GMT

We are running Oracle 8i on some of our database servers. The data resides in a VA7400. We have had performance problems that are indirectly related to when we create several 200MB archived logs in an hour.
Currently we have our Oracle data on 1 controller and the indexes on another controller. The log files are primarily on the controller the Oracle data resides.
Does anyone have any suggestions on arrange the logs in the array to get the best performance?

Re: Oracle Logs

Roger_22 — Fri, 18 Apr 2003 13:33:06 GMT

There are many factors that can influence the performance of the VA, particularly when you change the LUN configuration and workload.

But first, the basics; Firmware must be HP18. Normal resiliency, Prefetch enabled, Optimize Enabled. Use the armdsp ???a to verify the current settings. Use armmgr to change. Contact HP for the firmware upgrade.

Write again when the array is in this configuration if you still have a problem. Include the output from armdsp -a.

Re: Oracle Logs

Michael Steele_2 — Fri, 18 Apr 2003 14:44:43 GMT

Performance problems develope when RG's get a request for a disks that it doesn't own and has to query the other RG.

Think about this configuration:

Creating two seperate luns on each redundancy group, RG1 and RG2, then create two seperate PV's and add them to one VG and stripe your LV's across each PV using pvlinks. Make sure you have the primary LVM PV-Link going to its own controller and RG and you'll get the best performance.

Re: Oracle Logs

Christopher Caldwell — Fri, 18 Apr 2003 14:57:46 GMT

The VAs aren't the same as the XPs, but [from Oracle's perspective] the design and layout ideas should be the same:

Read
http://otn.oracle.com/deploy/availability/pdf/SAME_HP_WP_112002.pdf

NUTSHELL:
Alternate striping across all the controllers for all of the data; stripe all of the data across all of the LUNs.

Re: Oracle Logs

Thomas Maddox — Fri, 18 Apr 2003 16:18:09 GMT

Roger, you can verify these settings in my attached file. I am asking this question because when my system experiences high abort sequences on one controlller, we notice that the Oracle logs have high disk latency. The Oracle logs are on another controller.

Re: Oracle Logs

Michael Steele_2 — Fri, 18 Apr 2003 18:10:17 GMT

"...The Basics..., Firmware must be HP18. Normal resiliency, Prefetch enabled, Optimize Enabled..."

Which you have from your attachement:

Product Revision:_____________________HP18

Simplified Resiliency Setting:________Normal Performance (Default)

Enable Prefetch:______________________True

Enable Optimize Policy:_______________True

Trust me when I say "..better to have one big LUN in vg than many little LUNs...."

Try this link:

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x141c31ec5e34d711abdc0090277a778c,00.html

Re: Oracle Logs

Thomas Maddox — Fri, 18 Apr 2003 19:43:01 GMT

Thanks for referencing my own old call. I have been in touch with HP a few times on my VA7400 performance. They are fine with the number of LUNS I have. We just set the queue depth on LUNS as needed to adjust performance in some cases. Actually they found three issues. First, is the amount of space used in redundancy group 2. Second, the amount of free space in redundancy group 1. Third, we reduced the queue threshold from 72 to 60 on the array. The second answer sounds wild, but we went as high as we could with HP on the disk array.

Thanks a lot everyone. Especially Roger.