I. ASM-related concepts
1. What is ASM (Auto Storage Management)
Abbreviation for automatic storage management, a tool for managing disks
Ability to implement striped, mirrored data files, recovery files, etc. across multiple physical devices
The file is distributed across all disks in the disk group by allocation Unit AUS (allocation units), and ASM uses indexing technology to track the location of each AUS
Support for dynamic increase and decrease of online disks, and AUS will automatically regain dynamic distribution once the disk has changed
Support RAC Cluster technology, run an ASM instance on each node, and realize point-to-point communication between ASM instances
is a pure software level implementation, Third-party RAID work at the volume level, using a uniform stripe size, ASM can work at the file level, different files can be used
The different stripe size
The advantages of 2.ASM
Disk increase: Increasing disk becomes very easy. No downtime is required and the file area is automatically reassigned.
I/O allocation: I/O is automatically distributed across all available disks without human intervention, thereby reducing the likelihood of hot spots appearing.
Stripe width: Segments in redo log files can be subdivided (K for faster transfer rates), and for data files, bands are slightly larger (MB, to one-time
Transfer a large number of data blocks).
Buffering: ASM file system does not buffer, direct input/output.
Mirroring: If hardware mirroring is not available, software mirroring can be easily established.
Core asynchronous I/O: implementation of core asynchronous I/O requires no special settings and eliminates the need to use raw or Third-party file systems (such as Veritas Quick I/O)
3.Oracle RAID level for common data files
SYSTEM, undo table space is usually placed on a RAID 1 volume
Online redo logs are usually placed on a RAID 0 volume
Control files are usually placed on a raid 0+1 volume
Data files are typically placed on a RAID 5 volume
4.ASM architecture (mainly composed of ASM instances and ASM disk groups)
The use of ASM requires the creation of an ASM instance for managing ASM disk Groups
The ASM Disk group consists of an ASM disk that can contain one or more ASM disks
The ASM disk can be an actual disk, a partition of a disk, or a logical volume managed by LVM, but must be an unformatted raw device
The size of the ASM disk group is the sum of the size of the ASM disk within the group, and the usable capacity varies according to the fault tolerance level
5.ASM examples and RDBMS examples
The ASM instance is similar to a common database instance (RDBMS) and is also made up of the SGA and a stack of background processes, with the SGA only up to MB for most systems
The Largepool in the ASM instance is used to store the extent map, which calculates the size of the largepool based on the size of the database, usually 100GB size requires 1MB storage extent map
A typical RDBMS instance locates a data file and opens its data file, whereas in a database that uses ASM storage, the work is taken over by an ASM instance, which is used by the ASM instance to locate and
Manage ASM disks, disk groups.
The ASM instance has a background process similar to a common RDBMS instance, such as SMON,PMON,LGWR,DBWR,CKPT, and adds two new processes, one is Rbal, the other is ARBN
Rbal: For coordinating and managing the dynamic balance between disk groups
ARBN: Can be used for more than one, to complete the movement of AU
The ASM instance simply locates and manages the ASM disk, cannot load or open the database, and therefore cannot read the data dictionary information
An ASM instance usually has a parameter that can be started or closed only by using the password file or operating system authentication as SYSDBA or Sysoper.
An ASM instance can serve multiple RDBMS, creating multiple ASM instances on a single host, but typically using an ASM instance on a single host
If an instance serves multiple RDBMS, it is recommended to separate the ASM Asm_home from the DB Oracle_home for future upgrades and maintenance
Creation and configuration of ASM disk groups after the ASM instance starts
An RDBMS instance using the ASM disk is the same as a common RDBMS instance, but there are two more background processes, RBAL,ASMB.
ASMB uses a server process such as ORACLE+ASM1 to establish a common RDBMS instance to an ASM instance session and to pass information about the disk file
Rbal to open an ASM disk that is positioned through an ASM instance
Note: The ASM instance does not replace an RDBMS instance to read or write data files, which are delivered directly to the ASM disk and RDBMS instances, and ASM instances are used only for locating data files
The ASM disk, disk area, and related information that you want. Assuming that the new data file is added, the RDBMS tells the ASM instance to create the data file, and the ASM instance allocates the disk area
(extent) and create a disk map address to return to the RDBMS instance, the RDBMS writes the data directly to the disk group.
File types stored in the 6.ASM instance
Control file parameter file online log file log file data file Temp file
Rman backup set, image Copy control file backup set DataPump dump files, change tracing file
II. Create ASM instances (below VMware + RHEL 5.4 + Oracle 10g R2 Demo)
1. Install ASM Package
Using the ASM instance, you need to download the ASM package required on the Oracle website. Download path: Asmlib download
Note Download support for the kernel of the package, divided into two parts, one is the library and Tools, and the second is drivers for kernel
As my Linux kernel is:
[Root@oradb ~]# Uname-rm
2.6.18-164.el5 i686
then download the corresponding package as:
Library and Tools
oracleasmlib-2.0.4-1.el5.x86_64.rpm
oracleasm-support-2.1.3-1.el5.i386.rpm
Drivers for Kernel 2.6.18-164.el5
oracleasm-2.6.18-164.el5xen-2.0.5-1.el5.i686.rpm
oracleasm-2.6.18-164.el5debug-2.0.5-1.el5.i686.rpm
oracleasm-2.6.18-164.el5pae-2.0.5-1.el5.i686.rpm
oracleasm-2.6.18-164.el5-debuginfo-2.0.5-1.el5.i686.rpm