I'm preparing to migration of my site of plesk/nginx to Cyberpanel/OLS. For last two weeks I was checking dozen of different configurations (Centos 7, Centos 8, Ubuntu 20.04) and comparing with niginx/Apache. I've decided to settle on Centos 8, but have noticed some repeatable issues:
1. Redis server is not installed, despite checking this option in install script. It has to be installed manually
2. Some PHP extensions are marked as not installed, while they are installed physically, like lsphp74-mcrypt, lsphp74-zip or lsphp74-pecl-redis
3. Service status is showing ProFTPd service as stopped, while it is running
4. Regardless of which option is enabled in OLS, brotli compression seems to not run
5. It is probably my fault - I wanted to check Enterprise version, how is handling brotli. I've put wrong license key and whole configuration collapsed. Switch to enterprise failed, but OLS crashed and couldn't be started again from panel. After starting from command line, configuration was corrupted and cache wasn't working.
6. Last point is very nimble, but it was noticed by some users in previous posts. After setting up LS cache plugin in Wordpress, first read is quite slow, while second is blazing fast. By saying quite slow I mean compared to pure OLS without cache plugin. If cache is warmed, then GTMetrix/Pingdom/Pagespeed is flying like crazy. This is biggest puzzle for me.
I'm desperate to reach stable configuration, as all tests indicate consistent and better pagespeed results then nginx with rocket cache plugin.