RAC Environment:
Oracle Database: ORACLE11GR2
Operating system: rhel6.4_x86_64
Failure phenomena:
1. The process and instance of Oracle associated with ASM for its one node fails to start;
View method: Ps-ef|grep Ora_
Ps-ef|grep Asm_
CRSCTL Stat res-t (see only Oracle Instances and ASM instances launched on Node 1)
2, the ASM Disk device label is not visible on node 2;
View method: Oracleasm listdisks (no content)
ll/dev/oracleasm/disks/* (no content)
3, on Node 2 scan ASM disk times the following error;
View method: Oracleasm scandisks
Reloading Disk Partitions:done
Cleaning any stale ASM disks ...
Scanning system for ASM disks ...
Instantiating disk "VOL4"
Unable to instantiate disk "VOL4"
Instantiating disk "VOL1"
Unable to instantiate disk "VOL1"
4, view oracleasm log information prompt permission problem;
View method: Cat/var/log/oracleasm
Analysis Process:
In order to eliminate the interference of other factors, I restarted the Node 2 of the operating system, the restart after the ASM disk on Node 2 still does not load properly, by viewing the Oracleasm log still prompt is a permission problem, the bold conjecture, there may be the following reasons for ASM
The disk does not load properly:
1, ASM related RPM package installation is not normal, such as version does not correspond;
2, operating system-related security services caused;
3, the system restarts when the rule rules are not properly configured, resulting in the ASM disk does not load properly;
Exclude:
1, Rpm-aq|grep oracleasm need to install 3 packs (kmod-oracleasm oracleasmlib oracleasm-support)
2. Service iptables status (iptables services off)
Getenforce (View status enforcing, stating that the SELinux service is not off)
3. This instance does not use raw bare devices, so no rule rule is configured to load at startup
Processing method:
Because the SELinux Service for Node 2 is not turned off, the SELinux service is turned off
Setenforce 0
Vi/etc/sysconfig/selinux
Change the selinux=enforcing to selinux=disabled so that restarting the server does not start the SELinux service
After you close the SELinux service, perform an ASM disk scan operation to properly identify the ASM disk.
Oracleasm Scandisks
ASM disk Start failure location and recovery for a RAC node