Howard J. Rogers wrote:
>Daniel Morgan wrote:
>
>
>
>>Hari Om wrote:
>>
>>
>>
>>>I am using Oracle 9.2.0.1 on IBM AIX 5.1L
>>>I have come along with the following disk layout for my Database...can
>>>anyone point out if there would be any issues:
>>>
>>>hdisk0(18GB) + hdisk1 (18GB) --> These are MIRRORED. It contains OS +
>>>Oracle SW + Control file copy 1
>>>
>>>hdisk2(36GB) + hdisk3 (36GB) --> These are MULTIPLEXED by Oracle. It
>>>contains Redo Logs.
>>>
>>>hdisk4(36GB) + hdisk5 (36GB) --> These are MULTIPLEXED by Oracle. It
>>>contains Archive Destination.
>>>
>>>hdisk6(36GB) + hdisk7 (36GB) --> These are MIRRORED. It contains
>>>SYSTEM Tablespace.
>>>
>>>hdisk8 + hdisk9 + hdisk10 + hdisk11 + hdisk12 + hdisk13 --> Each are
>>>72 GB. These are MIRRORED and STRIPPED (1+0)--> These contain Oracle
>>>DATA fiels + Control File copy
>>>
>>>Can anyone point out any issues in the above layout....?
>>>Would it fine if we don't MIRROR REdo Logs?
>>>
>>>THANKS!
>>>
>>>
>>>
>>>
>>Mirroring something that is already mirrored by Oracle is creating a
>>redundant redundancy.
>>Given you have 5L I would make sure hdisks8 through 13 are RAW: No file
>>system.
>>
>>
>
>
>Oracle doesn't mirror. It multiplexes. And doing both is highly recommended.
>Hardware mirroring protects you against hardware failure; Oracle
>multiplexing protects you against software problems.
>
>And Hari asked this exact same question several weeks ago, and as I told him
>then, the real problem with this layout is the ridiculous waste of space
>that is associated with having two 36GB drives storing just the SYSTEM
>tablespace.
>
>And the other thing I repeatedly told him back when he originally asked was:
>if he's going to stripe and mirror all his other data files, what makes him
>think SYSTEM is so special it needs a disk+mirror all of its own?
>
>HJR
>
>
You are correct. I forgot that Hari seems to have a compulsive need to
re-ask questions as though they
were never answered. But I disagree about mirror what has been
multiplexed. I can't think of a possible
failure that would be better solved by having 4 control files than 2.
The issue is eliminating a single point
of failure. If you had two corrupt control files ... your hardware
mirroring would just give you mirrored
corruption.
--
Daniel Morgan
http://www.outreach.washington.edu/ext/certificates/oad/oad_crs.asp
http://www.outreach.washington.edu/ext/certificates/aoa/aoa_crs.asp
damorgan_at_x.washington.edu
(replace 'x' with a 'u' to reply)
Received on Mon Oct 20 2003 - 17:31:28 CDT
