Niall Litchfield wrote:
> "DA Morgan" <damorgan_at_x.washington.edu> wrote in message
> news:1102914028.50143_at_yasure...
>
>>Howard J. Rogers wrote:
>>
>>
>>>>You don't want to use raid 5.
>>>
>>>He might do.
>>>
>>>HJR
>>
>>Source:
>>http://download-west.oracle.com/docs/cd/B14117_01/server.101/b10752/instance_tune.htm#23421
>>Quote:
>>"Move the redo logs to faster disks or a faster I/O subsystem (for
>>example, switch from RAID 5 to RAID 1)."
>
>
> as a way to address log file sync waits.
>
> btw Amazingy the advice in this segment of the docs doesn't mention looking
> at whether the application is over committing or not.
>
>
>>Source:
>>http://download-west.oracle.com/docs/cd/B14117_01/server.101/b10752/iodesign.htm#26022
>>Quote:
>>"Avoiding the use of RAID 5 for redo logs."
>
>
> This is a wonderful piece of advice actually, I'll point folk at it in my
> bog later. What does it say
>
> 1. determine that your io bottleneck really is an io bottleneck.
> 2. determine where it lies
> 3. *if* it lies with redo take appropriate action.
>
> which is a little different from your quote
>
>
>>Source:
>>http://download-west.oracle.com/docs/html/B13831_01/tuning.htm#sthref531
>>Quote:
>>"Striping data across a number of disks is one example of a redundant
>>array of inexpensive disks (RAID). There are several different types of
>>RAID, also referred to as RAID levels, ranging from high performance to
>>high reliability. The three most common RAID levels in Oracle Database
>>installations are RAID-0, RAID-1, and RAID-5. Descriptions of each RAID
>>level follow Table 7-2, which shows each level's read and write penalties.
>>
>>RAID Level Read Penalty Write Penalty
>>0 (Disk Striping) 1:1 1:1
>>1 (Disk Mirroring) 1:1 2:1
>>0 + 1 1:1 2:1
>>5 (Distributed Data Gathering) 1:1 4:1"
>>.. and ..
>>"The write penalty of 4:1 results from 2 reads and 2 writes during parity
>>calculation."
>
>
> It does rather miss out the cost factor a bit though. RAID10 is better, but
> it isn't cheaper (or before folk jump on me as much more expensive as raid
> vendors say).
>
>
>>Source:
>>http://download-west.oracle.com/docs/cd/B10501_01/server.920/a96533/iodesign.htm#22610
>>Quote:
>>"RAID 0+1: Striping and Mirroring
>>
>>This level combines the technologies of RAID 0 and RAID 1. It is widely
>>used because it provides good reliability and better read and write
>>performance than RAID 1."
>>and
>>"RAID 5 striping is similar to striping in RAID 0 ... while write
>>performance can suffer. This configuration might not be ideal for
>>write-intensive applications."
>>
>>You are, of course, correct as always and I have obviously
>>misunderstood. Thank you, in advance, for correcting both me and Oracle.
>
>
> The Oracle quotations you provide address specific problems (or the benefits
> and costs of different raid configurations) - they do not amount to a *do
> not use RAID5* directive from Oracle. They amount (in my view anyway) to a
> list of the advantages and disadvantages of various IO platforms and the
> strategies to deal with specific problems. Howard used the word *may* and I
> think that is exactly correct RAID 5 may be appropriate or it may not.
Thanks Niall. At least *someone* got the point.
Regards
HJR
Received on Mon Dec 13 2004 - 01:31:09 CST
