Installation Using the CLI

This section describes how to install the fault diagnosis component using the CLI. The fault diagnosis component of 5.0.0.2 or later can be installed using a tool. For details, see Installation Using MindCluster Ascend Deployer.

Prerequisites

Ensure that the network is available before the installation.
The installation logs are not dumped. Before the installation, pay attention to the remaining drive space.

Procedure

Change umask to 027 by referring to Setting umask.
Upload the software package obtained in Software Package Acquisition to any directory (for example, ~/software) in the environment.
Decompress the package in the directory where it is stored.
```
unzip Ascend-mindxdl-faultdiag_{version}_linux-{arch}.zip
```

Install the component.

pip3 install ascend_faultdiag-{version}-py3-none-linux_{arch}.whl --log ~/.ascend_faultdiag/install.log

Change the directory permission.

chmod 700 ~/.ascend_faultdiag
chmod 600 ~/.ascend_faultdiag/*.log

Check whether the software is successfully installed.
```
ascend-fd version
```
Command output:
```
ascend-fd ${Version}
```

The default path for storing run logs of MindCluster Ascend FaultDiag is $(HOME)/.ascend_faultdiag/RUN_LOG/ directory. Each time the ascend-fd parse or diag command is executed, a folder containing a timestamp and a random number is generated. In the folder, run logs are flushed to drives by PID. If the number of folders in the RUN_LOG directory exceeds 100, 20 folders will be deleted by time. If the number of folders does not exceed 100 but the total memory size of files exceeds 100 MB, previous files will be deleted by time, and files of up to 80 MB will be retained.
The default path for storing operation logs of MindCluster Ascend FaultDiag is ${HOME}/.ascend_faultdiag/ascend_faultdiag_operation.log.
The size of a log file cannot exceed 10 MB. If the size of a log file exceeds 10 MB, another log file is automatically dumped. The number of log files with the same PID cannot exceed 10. If the number exceeds 10, the earliest log file is automatically overwritten.
If you need to customize a path for storing logs, see Customizing the Home Directory of MindCluster Ascend FaultDiag.
The device resource analysis and network congestion analysis modules of MindCluster Ascend FaultDiag depend on scikit-learn, pandas, NumPy, and joblib. To use these functions, ensure that the dependencies meet the version requirements: scikit-learn ≥ 1.3.0, pandas ≥1.3.5, NumPy ≥1.21.6, and 1.2.0 ≤ joblib < 1.5.0.

Parent topic: Installing Components