Solutions to Oracle faults on database servers running sun-Cluster

The database server and the Primary Application Server are deployed on two servers to achieve suncluster. In this case, troubleshooting is difficult.

Because the Oracle service is monitored by the cluster software, the cluster software will try to switch the shared disk array to another node when the service fails,

Start the Oracle service on this node. Oracle cannot be started successfully due to Oracle Database File Failure. When the attempt to start the service times out,

The cluster tries to start the Oracle service on the original node, resulting in the switch between the shared disk array and the service on the two machines.
Oracle services and related disk files are always in motion, which makes it difficult to view the database logs and troubleshoot database faults.

Therefore, stop the cluster to troubleshoot database faults.

It is very troublesome to stop the suncluster dual-host software. After the system is stopped together, restart the system with boot-X in the OK State.

Therefore, I decided to suspend the monitoring of the Oracle service from the Resource Group so that the cluster no longer switches due to Oracle service problems.

1. view the resource name in the Resource Group

Scstatus-G

2. Stop Oracle resource Switching
Stop resources:

Scswitch-n-J Oracle-lsnr-Rs

Scswitch-n-J Oracle-server-Rs

Command Format: scswitch {-E |-n}-J resource [,...]
Whether or not the database fails to be started will no longer lead to dual-host switchover. You can troubleshoot database faults.

2. Database troubleshooting

When starting the database, the following error is reported:
SQL> alter database open;
Alter database open
*
Row 3 has an error:
ORA-01113: File 1 requires media recovery
ORA-01110: Data File 1: '/Oracle/oradata/dbnms/system01.dbf'

To restore the database, run the following command:
SQL> recover Database
Media recovery is completed.

SQL> alter database open;

The database has been changed.

Open the database safely.

Cause Analysis: Generally, when the system01.dbf file is corrupted and the database is in non-archive mode, the database cannot be saved. Because the database redolog is divided into four groups, the change information of system01.dbf is not overwritten in redo. log. Therefore, during recover, the transaction data is found in redo04.log and restored successfully.

Therefore, to improve the database reliability, the database must adopt the archive mode. If the archive mode is not available, you can also increase the number of redo. log files to prevent the redo. log information from being quickly overwritten.

Appendix: Background log of alert_dbnms.log

Tue Feb 14   13 : 54 : 03  2012 
Alter database recover Database
Tue Feb 14   13 : 54 : 03   2012 
Media recovery start
Parallel recovery started 16 Processes
Tue Feb 14   13 : 54 : 12   2012 
Recovery of online redo log: Thread 1 Group 4 Seq 325853 Reading mem 0 
Mem # 0 Errs 0 :/Oracle/oradata/dbnms/redo04.log
Tue Feb 14   13 : 54 : 45   2012 
Media recovery complete (dbnms)
Tue Feb 14   13 : 54 : 47   2012 
Completed: Alter database recover Database
Tue Feb 14   13 : 54 : 59   2012 
Alter database open
Tue Feb 14   13 :55 : 00   2012 
Beginning crash recovery 1 Threads
Parallel recovery started 16 Processes
Tue Feb 14   13 : 55 : 00   2012 
Started redo Scan
Tue Feb 14   13 : 55 : 02   2012 
Completed redo Scan
 274807 Redo blocks read, 0 Data blocks need recovery
Tue Feb 14   13 : 55 : 02   2012 
Started redo application
Thread 1 : Logseq325853 , Block 183394 
Tue Feb 14   13 : 55 : 02   2012 
Recovery of online redo log: Thread 1 Group 4 Seq 325853 Reading mem 0 
Mem # 0 Errs 0 :/Oracle/oradata/dbnms/redo04.log
Tue Feb 14   13 : 55 : 04   2012 
Completed redo Application
Tue Feb 14   13 : 55 : 04   2012 
Completed crash recovery
Thread 1 : Logseq325853 , Block 458201 , SCN 16922087765 
 0 Data blocks read, 0 Data blocks written, 274807 Redo blocks read
Tue Feb 14   13 : 55 : 05   2012 
Thread 1 Advanced to log Sequence325854 
Thread 1 Opened at log Sequence 325854