Problem background and phenomenon
The installation cluster progress bar has no start time to fail directly:
Possible causes and solutions
Possible causes of database anomalies:
[Problem Cause 1]:
"/etc/profile" has a readonly environment variable that causes the exit when executing source /etc/profile.
[Solution 1]:
Execute vi /etc/profile, delete or comment the readonly line.
[Problem 2]:
The "/home/omm/.profile" permission is incorrect, causing the Gaussian database to be abnormal.
[problem phenomenon]
Observe the "$OMS_LOG_PATH/scriptlog/ha/scriptlog/ha.log" log, which prompts "java: command not found".
【identify the problem】
The general prompt "java: command not found" is caused by the WCC decryption database password when executing the dbfunc.sh file. The error is caused by not recognizing the java command.
There are two aspects to identifying this issue:
1. Whether the jdk path is included in "/home/omm/.profile".
2. Is it possible to access "/home/omm/.profile" under the ommdba account, and it is confirmed that the "/home/omm" directory is 700 permissions in the environment, and "/home/omm/.profile" cannot be accessed under the ommdba account. ".
[positioning conclusion]
The "/home/omm/.profile" permission is incorrect, causing a Gaussian database exception.
[Solution 2]:
If the jdk path is not included in /home/omm/.profile, increase it.
If the ommdba account cannot access "/home/omm/.profile", please modify the permissions.
