From: "Goulet, Richard" <Richard.Goulet@parexel.com>
To: "anthony.ballo@onecall.com" <anthony.ballo@onecall.com>,
 "oracle-l@freelists.org" <oracle-l@freelists.org>
Subject: RE: Virident FlashMAX SCM
Thread-Topic: Virident FlashMAX SCM
Date: Wed, 6 Jun 2012 15:01:37 +0000
Message-ID: <7642C3D639A75E4284658CAA38BF60A409D529C3@US-BOS-VEX001.eu.pxl.int>
References: <CBF3A4E7.392C7%anthony.ballo@onecall.com>
In-Reply-To: <CBF3A4E7.392C7%anthony.ballo@onecall.com>
Accept-Language: en-US
Content-Language: en-US
Content-type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
MIME-Version: 1.0
Sender: oracle-l-bounce@freelists.org
Errors-to: oracle-l-bounce@freelists.org
Precedence: normal
Reply-To: Richard.Goulet@parexel.com

Darn that myth is still running around?  No you're not missing anything, but the author certainly is.  Goes along with the individual who told me that you can drop the system tablespace.  Yes you can, but your db crashes in about 0.5 mseconds thereafter.

Richard Goulet
Senior Oracle DBA/NA TEAM Lead
PAREXEL International
T 978.495.4127
 
This communication, including any attachments, is intended only for the person or entity to which it is addressed and may contain confidential material. Any review, retransmission, distribution or other use of this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please destroy any copies, contact the sender and delete the material from any computer.  Thank you.


-----Original Message-----
From: oracle-l-bounce@freelists.org [mailto:oracle-l-bounce@freelists.org] On Behalf Of Anthony Ballo
Sent: Tuesday, June 05, 2012 2:59 PM
To: oracle-l@freelists.org
Subject: Virident FlashMAX SCM

Was reading this white paper from Virident today:
Accelerating Oracle Databases and Reducing Storage Complexity and Costs. Virident FlashMAX SCM as Primary Storage.


Offered at: http://www.bitpipe.com/data/document.do?res_id37280483_183&srcP53040&asrc=EM_BRU_17570135&uid956921


They are making a point that switching from 8k block size to 4k would reap performance benefits:


When Oracle data is stored on FlashMAX devices, reducing Oracle database block size from the default value of 8192 bytes (8KB) to 4096 bytes (4KB) can provide substantial performance benefits in many applications. With HDDs, reading/writing 4KB takes essentially the same amount of time as 8KB as most of the time is spent on moving heads. In contrast, FlashMAX can perform 2x the amount of IOPS with 4KB block size compared to 8KB block size, or the same amount of IOPS at lower latencies.


They offered the steps on how to change your block size:

You can set this parameter in several different ways:

 1.  By adding it to initORACLE_SID.ora file (or changing if the parameter already exists)  2.  By setting the parameter in the SPFILE:
    *   SQL>alter system set db_block_size@96 scope=spfile;
    *   SQL>shutdown immediate
    *   SQL>startup
 3.  By setting it on Initialization Parameters -> Sizing tab of the DBCA

Last I checked, you could only change the default block size at DB creation. Sure you can have multiple block sizes supported at the Tablespace level (must also specify cache sizes) but it's not as simple as changing a init parameter and bouncing your instance - otherwise, you will see:

ORA-00209: control file blocksize mismatch, check alert log for more info

Hard to take this white paper seriously or am I missing something?


Anthony


--
http://www.freelists.org/webpage/oracle-l


--
http://www.freelists.org/webpage/oracle-l