Step: Oracle 10g ASM On SLES9

1. Upgrade the SLES9 kernel and install the asm package. Ensure that the version is consistent.
2./etc/init. d/oracleasm configure the asm driver user as the oracle dba group and start automatically.
3. Prepare the raw disk to use fdisk for partitioning. If it is not formatted as any file system, initialize it by dd)
4. Create an asm Disk
Example:/etc/init. d/oracleasm createdisk d1/dev/sdb1
5. delete an asm Disk
/Etc/init. d/oracleasm deletedisk d1 (do not delete all disks in use; otherwise the database will crash)
6. Use dbca to create a database
Select asm, create diskgroup, select disk, and the disk starts to be in the Candidate state. After joining the disk group, it changes to Provisioned. If it is in the Member State, it indicates it was a Member of a disk group, to use this command, run the dd command to clear the existing ASM information on the disk. Syntax: dd if =/dev/zero of =/dev/sdd1 bs = 8192 count = 100 you can set the parameter: power limit 1-11). The smaller the value, the smaller the I/o bandwidth used by Asm for rebalance asm configuration change, the slower the process is, and the smaller the impact on system performance. Redundancy generally selects external, that is, it does not consider Redundancy. If normal is selected, it is mirror. At least one FailGroup must be set to High, which is triple mirror. Three times the image, three failgroups are required, the FailGroup actually divides the disk of the disk group into several groups and mirrors each other. In addition, discoverString is generally ORCL :*. Next, dbca will create an ASM instance and a database instance.
7. Modify/etc/oratab to change the N of the ASM and database instance to Y.

A bug about Startup Bug: 3458327-Automatic Startup On Reboot Fails When Database Uses ASM ORA--29701: unable to connect to Cluster Manager's current version 10.1.0.3.0. solution: (1) Change the/etc/inittab file to h1: 35: respawn:/etc/init. d/init.css d run>/dev/null 2> & 1 cut to l3: 3: wait:/etc/init. before d/rc 3

(2) modify the/etc/init. d/oracle file and add the following script before Starting the Database + Starting Database (s )...

Echo "Waiting css daemon start for 300 seconds! "Let count = 0 while [$ count-lt 30]; do sleep 10 let count = count + 1 let flag = count % 5 let seconds = $ count * 10 if [fag-eq 0]; then echo-n $ seconds else echo-n ". "fi done is based on seconds, the system initwill block init.css d for 5 minutes. "INIT: Id" h1 "respawning too fast: disabled for 5 minutes" (3) Run s21init.css d-> S13init.css d S13oracle  S21oracle in rc3.d and rc5.d

8. Manage the ASM disk group
(1) web
Check ORACLE_SID, export | grep ORACLE_SID on the database server to see if the database to be managed is not export ORACLE_SID = "xxx", then write down the url in emctl start dbconsole, and manage it in the browser on the client. log On As sys dba, select Disk Groups for Storage for administration, and select the Disk Groups for Storage when creating the database with the sys logon password) then we can manage asm: Create/delete/Mount/detach a Disk group, add/delete disks, modify and configure Disk Discovery Path, Rebalance Power, and Auto Mount Disk Groups.
(2) sqlplus command line
On the database server, you must first use export ORACLE_SID = + ASM, and then use sqlplus sys/passwd as sysdba to log on to create a disk group, add or delete disks of the disk group, and perform other operations. Alter diskgroup dgName drop disk diskname;
Alter diskgroup DG1 add disk 'orcl: D4 '; // it may be time-consuming because data needs to be transferred
Create diskgroup dg4 external redundancy disk 'orcl: 1' // CREATE a DISK group
Alter system set ASM_DISKGROUPS = 'dg1 ', 'dg2', 'dg3', 'dg4' // SET Automatic mount