RE: ASM of any significant value when switching to Direct NFS / NetApp / non-RAC?

From: Taylor, Chris David <ChrisDavid.Taylor_at_ingrambarge.com>
Date: Wed, 8 Aug 2012 07:30:07 -0500
Message-ID: <C5533BD628A9524496D63801704AE56D75B2FD239B_at_SPOBMEXC14.adprod.directory>



Looks like I may have perpetrated a myth about ASM rebalancing to reduce "hot spots". It appears that Oracle ASM has no "intelligent" monitoring built in to reduce hot spots - rather, the way extents are allocated across the logical volume units *helps* reduce IO hot spots, but doesn't necessarily eliminate them and ASM doesn't actively rebalance based on IO performance.

Chris Taylor

"Quality is never an accident; it is always the result of intelligent effort."
-- John Ruskin (English Writer 1819-1900)

Any views and/or opinions expressed herein are my own and do not necessarily reflect the views of Ingram Industries, its affiliates, its subsidiaries or its employees.

-----Original Message-----
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Taylor, Chris David Sent: Wednesday, August 08, 2012 7:13 AM To: 'dananrg_at_yahoo.com'; 'oracle-l'
Cc: 'wts_at_maine.edu'
Subject: RE: ASM of any significant value when switching to Direct NFS / NetApp / non-RAC?

One thing that comes to mind is the automatic rebalancing to reduce "hot spots" - Not sure how that works exactly in practice though. Another thing that comes to mind is double striping (aka plaiding, stripe-on-stripe) - Think of a plaid shirt, you have stripes across and down - when striped in this manner, IO performance is really optimized - not sure how that relates direct NFS though. (Make several LUNS that are physically striped across multiple physical disks in the storage unit, present those to the host, then at the host create DISK GROUPS using each LUN - stripe across the LUNS).

Here's a few links:

http://bartsjerps.wordpress.com/2011/05/25/why-oracle-asm/
http://asktom.oracle.com/pls/apex/f?p=100:11:0::::P11_QUESTION_ID:30908256823200
http://h71028.www7.hp.com/enterprise/downloads/4AA0-9728ENW_ASM_EVA_Best_Practices_White_Paper_final%20121806.pdf

"The EVA configuration for OLTP testing shows an advantage toward a configuration that uses both ASM and EVA striping (double-striping). The advantages were more pronounced as the database size increased (2.5-TB RAC database with 1,000 concurrent users)."

Also, get with your storage vendor and discuss double striping and find out if they have any whitepapers discussing Oracle ASM with direct NFS.

Chris Taylor

"Quality is never an accident; it is always the result of intelligent effort."
-- John Ruskin (English Writer 1819-1900)

Any views and/or opinions expressed herein are my own and do not necessarily reflect the views of Ingram Industries, its affiliates, its subsidiaries or its employees.

-----Original Message-----
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Dana Nibby Sent: Wednesday, August 08, 2012 2:34 AM To: oracle-l
Cc: wts_at_maine.edu
Subject: Re: ASM of any significant value when switching to Direct NFS / NetApp / non-RAC?

Thanks Wayne and Martin. We've always had striping through our SAN. And so have never had to use ASM for performance/fault tolerance. When you say there is more to ASM than striping, what functionality is that? Martin, you mentioned storage migration as one item. Would love to hear more details and use cases for non-overlapping ASM functionality (non-overlapping with NetApp RAID DP appliance storage with DNFS).

Without ASM, I understand there's one less piece for the sysadmins/storage admins to manage. And so it reduces complexity. But I want to be sure removing ASM does not cause any loss in our ability to rapidly respond to customers. Given the separation of duties in our organization and the workload of the sysadmins/storage admins, this is key. The bottom line: would removing ASM make the customer less happy, more happy, or would the change likely be 100% transparent? Will we DBAs have to rely more upon our sysadmins / storage admins to get things done, e.g. add and grow datafiles, etc? If so, customer service will by definition suffer. Because that's one more layer to go through to get things done.

Another issue: if we rebuild all our instances without ASM, can we still take advantage of Oracle Restart? In other words, do we continue to perform Grid Infrastructure installs and then just ignore / not configure an ASM instance? Perhaps it's better for another post, but are there reasons *not* to use Oracle Restart for non-RAC 11g instances?

Since I imagine most shops use RAID storage (and increasingly sophisticated SAN/NAS devices), how popular is ASM? Oracle still seems to tout ASM as a Best Practice. But a best practice for whom and in what context(s)? For those using JBOD? Maybe I'm thinking of Oracle literature back from 10g or the 1st release of 11g. If ASM isn't on the path to deprecation in today's storage landscape, why not?



 From: Wayne Smith <wts_at_maine.edu>
To: oracle-l <oracle-l_at_freelists.org> Sent: Tuesday, August 7, 2012 1:34 PM
Subject: Re: ASM of any significant value when switching to Direct NFS / NetApp / non-RAC?  

If you are moving from unstriped/non-raid disk to a disk appliance, that probably means ASM striping support is of no interest going forward, but their is more to ASM than that.

That said, I believe ASM is an unnecessary complication for my standalone/non-RAC databases.  Your mileage will vary.

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




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




--
http://www.freelists.org/webpage/oracle-l
Received on Wed Aug 08 2012 - 07:30:07 CDT

Original text of this message