Troubleshooting: Talend products installation on RedHat/CentOS 8 fails

Before you begin

Follow this troubleshooting article if the installation of Talend products fails under this specific configuration:

You use RedHat 8 or CentOS 8.
SELinux subsystem is active (this is the default behavior).
You are installing Talend products on a second hard-drive or in a non-standard directory (/srv for example).

During the installation of the product, this error message appears:

Error: There had been an error. 
Error running /bin/systemctl start talend-tac-7.3.1 : 
Job talend-tac-7.3.1.service failed because the control process exited with error code.

When you execute the command journalctl -xe | grep talend-<product>-<version>.service (in this example: journalctl -xe | grep talend-tac-7.3.1.service), this is the result:

systemd : talend-tac-7.3.1.service: Failed to execute command: Permission denied
systemd : talend-tac-7.3.1.service: Failed at step EXEC spawning /mnt/opt/Talend-7.3.1/utils/srv_tac: Permission denied

Procedure

Check the file labels in the installation folder. Make sure usr_t is used. With SELinux, SystemD can only run scripts from files labeled usr_t.
In the following example, the /opt directory is compared to the /mnt/opt directory.
```
[root@centos8 ~]# ls -dZ /mnt/opt
system_u:object_r:mnt_t:s0 /mnt/opt
[root@centos8 ~]# ls -dZ /opt
system_u:object_r:usr_t:s0 /opt
```

If usr_t is not used, change the labels.

[root@centos8 ~]# semanage fcontext -a -t usr_t "/mnt/opt(/.*)?"
[root@centos8 ~]# restorecon -R -v /mnt/opt
Relabeled /mnt/opt from system_u:object_r:mnt_t:s0 to system_u:object_r:usr_t:s0
Relabeled /mnt/opt/test.sh from unconfined_u:object_r:mnt_t:s0 to unconfined_u:object_r:usr_t:s0

Results

You can now resume the installation. For more information, see RedHat documentation.