Limitations and Precautions for System Character Encoding

Hardware Requirements

Supported by all models.

License Requirements

The system character encoding function is not license-controlled.

When the Character Encoding Format Needs to Be Switched

The system supports GBK by default, meeting the input requirements of most languages including Chinese and English. The encoding format needs to be switched to UTF-8 only when characters (such as German and French) that GBK does not support are used.

In the Chinese environment, switching the UTF-8 encoding format is not recommended, the encoding format can be switched only on the English web UI or the CLI whose language mode is English.

Switching Impact

Once the encoding format is switched, the Chinese web UI and CLI are unavailable.
If the UTF-8 format is required, you are strongly recommended to switch the default format to UTF-8 before configuring any services. Otherwise, some configurations may not be automatically transcoded. In addition, if the device has user authentication-related configurations, the device will be restarted during transcoding, interrupting services.
You cannot switch the UTF-8 encoding format back to GBK, especially when the original configuration file has the SSL VPN configuration and the database cannot be rolled back. If the original configuration file does not have the SSL VPN configuration, you can restore the configuration file and database backed up earlier (the device automatically backs up the current configuration and database during encoding format switching). However, this operation is risky. Exercise caution before selecting this operation.
UTF-8 is a variable-length encoding format. Each character contains 1 to 6 bytes. After the encoding format is switched to UTF-8, some character strings entered in the previous GBK format may fail to be transcoded because their length exceeds that for UTF-8. During transcoding, delete or modify the overlong configuration according to prompts. The character string length prompted in CLI help information is the number of bytes. One English character occupies one byte; one Chinese character occupies two bytes. After the encoding format is switched to UTF-8, the length of one character may extend, but the prompt information is not dynamically changed. If the input character string is overlong, the system displays an error message.
For the execution of the language character-set utf-8 command, if the FW has user management and SSL VPN functions configured and the configuration saved in the database, such as the

user name, user group name, security group name, or SSL user role, contains non-ASCII characters that need to be converted, the device automatically restarts. After the restart, the conversion is successful. In this case, you are advised to perform

the conversion during off-peak hours.

Switching Limitations

For original configurations, the FW supports transcoding only from GBK to UTF-8. Before the encoding format is switched to UTF-8, the web UI delivers only GBK configurations. When the CLI is used to configure the device, an administrator may use different encoding formats on the terminal software. Such configurations are stored in the corresponding encoding formats on the device. The FW does not ensure successful transcoding from non-GBK to UTF-8.
As the configuration delivered through the web UI can only be in the GBK format, while the configuration delivered through the CLI can be in the UTF-8 format (depending on the CLI terminal), these two encoding formats may co-exist on the FW .In this case, the FW does not support transcoding.
The encoding format switching needs to be performed through the web UI or CLI. You are prohibited from adding the command line for switching the encoding format in the configuration file and then importing the configuration file to the device.
The device may be restarted during transcoding if it has user authentication-related or SSL VPN configurations. To ensure non-stop services, in hot standby networking, perform transcoding on the standby device first, manually switch the active and standby devices, and perform transcoding on the original active device (the transcoding command does not apply to hot standby). The procedures for the active/standby and load balancing networking modes are different. For details, see the configuration procedures.
The keywords configured in the content filtering feature do not support the UTF-8 encoding format. After the encoding mode of the device is switched to UTF-8, the device supports only ASCII characters, such as letters, digits, and symbols.
The file sharing configuration of SSL VPN does not support the UTF-8 encoding mode. If non-ASCII characters exist, garbled characters are displayed after the encoding mode is switched to UTF-8.

Impact on Interconnection

After the UTF-8 encoding format is used, note that the encoding format of the CLI terminal software must be set to UTF-8 as well. Otherwise, entered characters may be displayed as garbled characters.
The FW character encoding format must be consistent with that of the interconnected third-party system, such as the authentication server. Otherwise, the interconnection fails if there are non-ASCII characters.
When you import data to the FW through a CSV or TXT file, the encoding format shall be consistent with that of the FW. Otherwise, garbled characters may appear after the import.