Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Fundamental RAID ISSUES

Re: Fundamental RAID ISSUES

From: Jerry Gitomer <jgitomer_at_hbsrx.com>
Date: Fri, 18 Jun 1999 15:06:34 -0400
Message-ID: <7ke98e$dam$2@autumn.news.rcn.net>


When it comes to RAID there are three benefits:

        High reliability
        Low price
        High performance

Pick the two you want. I am not being facetious about this. There is a tradeoff. RAID 5 will give you low price and high reliability at the expense of performance. RAID 1+0 will give you high reliability and high performance.

There are three potential problems with RAID 5 on NT:

  1. Don't even consider using software RAID 5. I know it's cheap and you will certainly have a reliable system, but it is painfully sloooow.
  2. RAID 5 takes at least twice as long to write a record as it does to read one. The reason is that parity is calculated as the data is written and is then written on the next disk revolution. Depending on your RAID vendor, the characteristics of your drive, and perhaps even the direction the wind is blowing the time between writing the data and the parity can vary. Worst case is not being able to write for a full revolution which translates into 3 revolutions per write as compared to one per read.
  3. Even if you decide to use RAID 5 be sure to move your REDO logs off of the RAID 5. This is because your REDO logs are where the action is. Oracle is constantly writing to the REDO logs and only writes back to the database when it must.

regards

Jerry Gitomer


coakleyj_at_hotmail.com wrote in message <7kd6sg$ds7$1_at_nnrp1.deja.com>...
>A couple of fundamental questions regarding RAID 5 configurations for an
>Oracle OLTP application, involving 60% query, 40% writes approximately
(NT
>4)-1. In general, which is the better configuration (and why) - Six 9GB
>10K RPM disks or Three 18GB 10K RPM disks - both configured as RAID 5. (I
>presume the more disks the better to reduce disk I/O?). Is there an upper
>limit to the number of disks that can be configured as one RAID set
>(before management overhead becomes too large)?2. In terms of the amount
>of disk space "used-up" on the RAID parity - how does this differ in the
>above configurations? How is this calculated? - It seems to be a function
>of the number and size of the disks in the RAID set. Is there a formula?3.
>What are the implications of the "stripe width" when configuring the system
>- What's "normal" ? How does this interact with the O/S block size (if at
>all), and how do these settings interact with the Oracle block size
>setting? Which drives the Oracle block size - the stripe width or O/S block
>size?4. Previous questions and answers on this SIG have suggested that
>RAID 5 decreases write performance whilst improving reads. The reason given
>(I think) was that the write involved 2 writes really - one for the data
>and one for the parity. Does the same logic not apply to reads - i.e. with
>RAID 5 we need to perform 2 reads instead of one ? Hence reads are slower
as
>well??Can the system be configured to optimize reads over writes or vice
>versa?5. Finally, a licensing issue regarding NT4- Assuming we have 50
>concurrent Oracle users on an NT4 server. Given that the users are
operating
>client-server and only need "ping" capability to the NT server (i.e. the
>end-users are not set up as users on the NT server, nor do they have access
>to the disk drives), do we need to licence 50 users of NT or not? If not,
>how many?RegardsCoakleyj
>
>
>Sent via Deja.com http://www.deja.com/
>Share what you know. Learn what you don't.
Received on Fri Jun 18 1999 - 14:06:34 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US