Running status of Oracle RAC

Introduction: in the past few days, the RAC built on the two servers run a few extremely high-computing jobs, so I am very concerned about the performance of the database's Linux server. In Linux, I used the TOP command to monitor the overall consumption of the server's CPU and memory. I found that the CPU of a server only has 50% IDs, which indicates that only 50% of the remaining CPU is idle, the CPU of the other server is 97% id, which indicates that there is 97% idle. This status indicates that one of the two servers is overloaded and the other is not running at all, why is this happening? Should RAC be able to achieve load balancing ?? Follow these steps:

1. Run the following statements: srvctl status database-d sdwjsis: I

Instance sdwjsis1 is not running on node sdwjsis01

Instance sdwjsis2 is not running on nodesdwjsis02

Why is this problem? When the two database servers are running, how can no instances run on sdwjsis01 and sdwjsis02?

The initial analysis may be that when I perform some UPDATE operations on the Stored Procedure statement of the JOB, I restarted one of the Linux servers using the usual command shutdown immediate to shut down the database and then restart the database. This traditional restart process may cause the RAC to fail to start normally.

In this case, the Oracle Forum explains: because wedidn't start the database with SRVCTL and eventhough database was open manually usingsqlplus the cluster has no clue on it. that is to say, the old method cannot be used to restart the RAC database.

2. For the correct restart and shutdown methods, see:

1,

2. the restart and shutdown methods for Oracle 10g and 11G are different. For details, see: