RE: ASM IO distribution on a SAN

  • From: ryan_gaffuri@xxxxxxxxxxx
  • To: "Kurth, Michael J." <Michael.Kurth@xxxxxxxx>, <oracle-l@xxxxxxxxxxxxx>
  • Date: Thu, 13 Sep 2007 12:48:40 +0000

ok so I should create 1 ASM disk group for each LUN. 

I do not know alot about SANs. Is it possible to map a LUN to a partition or 
mount point? Is that how you make it visible at the Oracle level? 

Also, the SAME article by Juan Loazia talks about only using the outer half of 
disks. I am assuming this is something a SAN administrator can configure for 
me? 
-------------- Original message -------------- 
From: "Kurth, Michael J." <Michael.Kurth@xxxxxxxx> 

We use ASM here.
 
Each asm disk group consists of several asm disks.
Each asm disk is actually a lun from your SAN.
Each lun consists of 8 physical disks.
 
When you create a tablespace you specify an asm disk group to put the data.
 
So when you create a table, it gets striped very evenly across each of the asm 
disks.
 
The more asm disks in your disk group, the more physical disks your tables get 
spread across.
Mike Kurth 
Sr Database Administrator 
Wheaton Franciscan Healthcare 
400 West River Woods Parkway 
Glendale, WI 53212 
Phone: 414-465-4046 
Pager: 414-584-0042 
michael.kurth@xxxxxxxx 



From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On 
Behalf Of ryan_gaffuri@xxxxxxxxxxx
Sent: Wednesday, September 12, 2007 9:56 PM
To: oracle-l@xxxxxxxxxxxxx
Subject: ASM IO distribution on a SAN
 
The Oracle docs state that ASM can perform striping and mirroring. If I use ASM 
on a SAN how is it able to stripe? All it will see are mount points? My 
understanding is that it is possible create LUN(logical Units) and map them to 
individual disks or sets of disks and then possibly map those to mount points. 
Would this make it easier for ASM to handle IO distribution?
 
The project I am working will fully deploy with 5 billion inserts(all bulk 
loads using external tables) and have 4 million OLTP type queries/day (average 
query returns over 100 rows, but they are all index scans and the queries are 
not that complex). No updates or deletes. 
 
I am trying to find ways to optimize our physical IO. We are using RAC so if I 
need more CPU we can add nodes. 
Privileged/Confidential information may be contained in this message. The 
information contained in this message is intended only for the use of the 
recipient(s) named above and their co-workers who are working on the same 
matter. The recipient of this information is prohibited from disclosing the 
information to any other party unless this disclosure has been authorized in 
advance. 

If you are not intended recipient of this message or any agent responsible for 
delivery of the message to the intended recipient, you are hereby notified that 
any disclosure, copying, distribution or action taken in reliance on the 
contents of this message is strictly prohibited. You should immediately destroy 
this message and kindly notify the sender by reply E-Mail. Please advise 
immediately if you or your employer does not consent to Internet E-Mail for 
messages of this kind. Opinions, conclusions and other information in this 
message that do not relate to the official business of the firm shall be 
understood as neither given nor endorsed by it. 

Other related posts: