I know EMC has some recommendation for ASM configuration for Oracle 10G:
ASM disk group configuration guidelines• Use multiple disk groups – preferably a minimum of four, optimally five;
place the Data, Redo, Temp, Undo, and FRA/archive logs in different
(separate) disk groups
• Optimally, use RAID 1 for tier-1 storage
• Use external redundancy instead of ASM mirroring
• Configure disk groups so that each contains LUNs of the same size and
• Distribute ASM disk group members over as many spindles as is practical for
the site’s configuration and operational needs
From what I heard, Oracle 11G ASM has huge difference from 10G. Has EMC published any guideline for ASM configuration? Thanks.
I have checked Oracle 11G ASM document and found some recommendation emphasizing on hardware stripe size (64K, 128K, 256K) to match the ASM AU. FYI
Hardware RAID Striped LUNs
In the case shown in the slide, you have a one-terabyte database with a corresponding
two-terabytes flash recovery area. You have eight arrays of twelve disks, with each disk being 73 GB. You use RAID 0+1, a combination of hardware striping and mirroring, to mirror and stripe each partition, creating four logical units of 876 GB each. These can be defined as an ASM single disk group. ASM mirroring is not used.
Here, you define bigger LUNs, not restricted to the size of one of your disks. ASM striping spreads the allocation units across the LUNs. The RAID-level striping further spreads I/Os across the physical disks. By doing this, you achieve a better data distribution across all your disks, and you end up managing significantly fewer LUNs.
However, you must manipulate your storage in much larger chunks than in the previous configuration. To force the use of the faster region for data, an additional LUN would have to be created across a set of partitions of the disks resulting in two LUNs.
Note: The hardware stripe size you choose is very important. You want 1 MB alignment as much as possible, to match the ASM AUs. Selecting “power of two” stripe sizes
(64 KB, 128 KB, or 256 KB) is better than selecting other numbers. The choices offered by storage vendors for stripe size and partitioning depend on their storage array technology. These can create unnecessary I/O bottlenecks if not carefully considered.
I happend to work on a ASM configuration case recently and find the below recommendations for ASM 11G. Mostly are from Oracle directly, hope it's helpful for you guys.
To reduce the complexity of managing ASM and its diskgroups, Oracle recommends that generally no more than two diskgroups be maintained and managed per RAC cluster or single ASM instance
oDatabase work area: This is where active database files such as datafiles, control files, online redo logs, and change tracking files used in incremental backups are stored. This location is indicated by DB_CREATE_FILE_DEST.
oFlash recovery area: Where recovery-related files are created, such as multiplexed copies of the current control file and online redo logs, archived redo logs, backup sets, and flashback log files. This location is indicated by DB-RECOVERY_FILE_DEST.
For all 11g ASM/DB users, it best to create a disk group using 4 MB ASM AU size. Metalink Note 810484.1 covers this
ASM Best Practices for LUN :