On a fresh CentOS 8 / Virtualmin install, I’ve been trying to sort this error out since last night and reading and trying to sort it out for most of today. The only change is the socket number, which has changed at each attempt (something throwing up a new instance of MariaDB when I close an open one down?).
I can’t find a way in Webmin/Virtualmin to uninstall MariaDB or to install a fresh instance. Am I missing something?
So I’ve decided to start again. My VPS host provides a VM for reinstalling OSs so it’s not a big deal. The question is whether my habit of battening down a VPS as soon as I get it may be causing problems?
The first thing I do on a VPS is to change root password, create a new non-root user, allow that user shell access and disallow root from all direct external access. I then install Fail2Ban and either enable or disable SELinux (disable to install Virtualmin) prior to installing anything. I can’t see anything wrong with any part of this procedure, but I could easily be wrong.
Finally, the root user I’ve banned from direct access to the VPS appears to be the only user who can get access to Virtualmin. Is this considered to be secure?
I’m not sure about the causal connections here, happy to break it up into multiple threads if that helps.
Would somebody please point me in the right direction?