UCP not running, cannot connect to database

[cursor]

Today we had a nasty surprise. A freepbx 15 server that had been running for years was accidentally shut off and when we restarted it UCP refused to run. The error we can see in ucp_err.log is the following:

{ Error: Access denied for user ‘freepbxuser’@‘127.0.0.1’ (using password: YES) code: 1045 }

The stock install only makes user freepbxuser accesible from “localhost” so why would UCP want to use 127.0.0.1 for its host? We had to create another freepbxuser account with the same password and permissions but with host 127.0.0.1 so UCP could start. Why did this configuration change? Where is this configuration? Shouldn’t all freepbx apps use freepbx.conf in /etc to pull the database configuration?

Since UCP was not running we could not use Zulu so this prevented regular operations. I need to understand what happened in order to prevent something like this from happening in the future.

[jfinstrom (James Finstrom)]

From the error my guess would be that a password was changed in the database but not the config. In almost all cases 127.0.0.1 and localhost are able to be used interchangeably.

[cursor]

Nothing has changed in a long time. Like I mentioned the server has been working for years. The frreepbx.conf file is dated Nov 19 2020. MariaDB does differentiate 127.0.0.1 from localhost. The user created by freepbx has "localhost " as the permitted host so it will not allow 127.0.0.1 which is why we had to make the new user. Once the user was created UCP logged in and started working.