The inspection fails due to 100% CPU usage of controller A.
In the /other/cur_task_info log file of the primary controller, enter the cli keyword to search for information. If there are large amount of /ISM/cli/ismc processes, as shown in
the following screenshot, the fault is diagnosed.
a. Log in to the CLI using PuTTY or SSH, and run the
following command to go to the developer mode. The debugging password is debug@storage.
change user_mode current_mode user_mode=developer
b. Go to the minisystem mode.
c. Run the ps -C ismcli command to
query the PID of a CLI process.
d. Run the kill -9 PID command to end related
processes. Note: In the result queried in the preceding step, find all the PIDs
whose values in the TTY column are a
quotation mark (?) and end the processes in sequence.
e. Exit the CLI.
f. Use Toolkit to inspect
the device. The CPU usage check item passes the check and the alarm reporting
that the CPU usage of controller A reaches 100% is cleared.
1. Too many residual CLI processes cause the CPU usage of
controller A to be 100%. (There are 32 ismcli processes, and each process occupies 5.7% of the CPU.)
2. The residual CLI processes exist because the password
expires and the system displays a page of modifying the password. However, the
request is not processed. Each time you close the page, one CLI residual
process is generated. The system allows you to close the page for 32 times in
maximum. Therefore, there are 32 residual processes in maximum.
If this issue occurs on OceanStor 18000 V100R001C00SPC300,
end all processes whose TTY is a question mark .