I’ve used cyberpanel / vestaCP / hestiaCP / cPanel / directadmin and other panels for many, many, many years (maybe a decade for sure).
For avoiding a wall of nonneeded text, I will stick with only 2:
First, and most important thing:
Yes, OpenLiteSpeed or OLS itself will be ALWAYS faster than nginx + apache + php + mysql. Always. Tested, verified many many times. Out the box. Always. Yes, I was thinking vice versa for a long time, because I have dramatic experience with Nginx, but I miss OLS improving over years, and the Chinese guys did an amazing job in terms of performance. Seriously, without any jokes, and fakes. OLS really fast.
Let’s talk about average user setup like woocommerce / wordpress shit with 20-30 plugins.
What will be better in terms of performance under ideal conditions for people without experience at all?
I think, and I can prove any of my words, that out the box CyberPanel / OLS will be faster, better, and easier to use for common usage and much higher/faster results out the box.
I understand all of your love for hestiaCP / vestaCP, nginx, apache, whatever, you can close your eyes and blindly believe that all that they doing is “fakes”. But you can set up several instances as I did a few months ago with Vultr, and test it under real production. And yes, results surprising and not on the side of nginx + apache (hestiaCP).
So. Is OLS will be better? Almost for any bench/ your question yes.
BUT DO NOT RUSH TO SWITCH panels, webserver setups, whatever. Please keep reading.
Right now, as you see, even knowing that OLS faster, better, and bla bla bla, i still sitting on HestiaCP.
Why?
Why do I choose slower things for my needs?
The answer and reason/motivation are extremely simple: I WANT STABILITY.
Yes, I’ve been with vestaCP for a long time, then I switched to raw own nginx + php-fpm and other mods configs, but I’m too old for all of that shit, i know a lot of shit, hacks, updates, and other out the box tons really tons of problems and bugs with vestaCP, and sadly tons in hestiaCP…
But such mistakes are at least fixed on hestiaCP side, and at least the panel is stable.
I.E. you can not worry if you set up the panel, and move around 20-30 of your websites to hestiaCP, that one day you will not be even able to login to hestiaCP, due to fucking awful and stupid bugs full of extremely stupid chines bugs in OLS.
And here is started criticism of OLS / cyberpanel, that I’ve used for maaaaaany years too.
1.) I do not give a fuck if OLS faster, if OLS and their lsapi php shit can break one day without any notice, or provoke memory leaks, or has childish bugs in code, that provoke bugs with using ALL empty space / iops, etc. And you’re not even able to login to the server.
2.) I do not give a heck about features out the box in cyberpanel, ols, and so on if on one day, I woke up on 3 am and see that all of my websites were DEAD, and server usage was almost zero percent. You trying to find out what happened, and you can’t. Do you know why? Because the whole architecture of this piece of shit OLS is damn awful. Just damn awful. Who the heck wrote such logs format? Who the heck decides to trace THAT way issues/problems when something does not work?
I’ve made an instant backup, and tried for 1 hour to realize wtf is going on with my website, but i gave up and just setup a new instance and moved all of the websites to HestiaCP.
Then when my hobby projects back online with hestiaCP, i decide to back to not the closed case with cyberpanel and OLS.
And do you know what???
I spent around a WEEK, just fucking WEEK to trace, find, understand why the shit happened.
Do you know why? Because of luck of docs, luck of info, luck of everything. I was forced to use my system-admin background to trace DIRECTLY what software doing and debug it in gdb to understand WTF is happening, and why the shit is not documented?? Are you realize how crazy this shit is? Oh no, I think you don’t.
After a week I give up, I found tons of really awful things with OLS, with their official website uptime, stability, docs, docs quality, and so on, and i decide - ENOUGH, and destroyed instance with a snapshot of problematic OLS / cyberpanel. You know, i have a passion for debugging software, but this shit made me crazy. No way and no words to describe how things really bad.
So, let’s summarize:
1.) IS OLS faster than nginx + php-fpm / or apache + php-fpm or nginx + apache + php-fpm.
Yes, it is. Tested in December 2021.
2.) Will my WordPress website be faster with OLS + Litespeed plugin?
Yes, it will be faster, easier to manage, and better than any plugin in WP market. Tested all plugins.
And will require zero efforts, easy to use. Really.
3.) But if everything is so good, why are we talking about it on hesteiaCP forum?
Because things NOT so good. And the software is extremely bugged. Just extremely bugged, and not only software, EVERYTHING related to OpenLitespeed is extremely bugged. CyberPanel is extremely UNSECURE and BUGGED (hacked several times their Django shit on dozens of my setups and no word about their CVE / problems with security, wtf is that?).
Bugs can come AFTER update, after CyberPanel upgrade, after rebooting server, after enabling the plugin, after disabling plugin, after almost EVERYTHING. There is no QA at all.
All OLS things remind me of a student project. Tons of nice and fresh ideas, tons of really needed things, but when it comes to stability - there is no stability at all.
And yes, I’m tired of even writing the wall of text, because I can add tons of screenshots, videos, proofs about each of mine word/phrase. I will NOT apologize for that shit what people doing.
What about “benches”. here is one of the “fresh” ones, but in russian.
± adequate, ± same results as mine, but benched not only by bench tools but also mine sites were under production.
And last: as you see, I said ZERO words about Litespeed (enterprise level). This is totally different story, that has a dramatically lower rate of problems described in OLS above. (I do not really believe that OLS & Litespeed share the same codebase).
