LiteSpeed Instability after <3 hours

I’m having a lot of trouble with litespeed since I upgraded cyberpanel to 2.1.1 just a few days ago.

I used the cli to upgrade, and did a reboot just to be sure. I have to manually restart litespeed every 10min->3hours or all the sites on the server start going up and down.

I’m monitoring using Prometheus on a server from my home office(Which is whitelisted). Often by the time I try to pull up the website its already working fine, so the downtimes are brief. But they all load slowly once litespeed starts acting up.

I’ve been monitoring and restarting for days trying to check logs and figure out what’s going wrong.

I’ve tried connecting the cloud, and now cyberpanel isn’t using the right SSL, its defaulting to example.com now.

Please Help!

Well I’ve fixed the example.com problem with instructions from here: SSL still www.example.com - General Discussion - CyberPanel Community

I’ve gone thru and created a www DNS entry for each domain/subdomain on my server as insurance. Is worth noting I do not use this server for DNS at all. PowerDNS isn’t even enabled.

I too experience instability with LiteSpeed after upgrading CyberPanel to Ver 2.1.1 … Web server would lag to too long, and then stop responding. Sometimes it would fix by itself and other times I’d have to restart LiteSpeed, but regardless a short while later, webserver would go down again. :frowning:

As soon as I reverted to my backup (ver 2.0 build 2), then LiteSpeed was fully stable again.
The problem is … CyberPanel Ver 2.0 stopped issuing new SSL certificates from LetsEncrypt. Not sure what to do, but I can’t allow my webserver to keep going down with CyberPanel ver 2.1.1

Provider: Digital Ocean

System: Ubuntu 20.04

Installation option: upgrade from 2.03

OLS or Enterprise?
OLS

Installing from official server or mirror server?
official

Install Memcached extension for PHP?
no

Install LiteSpeed Memcached?
yes

Install Memcached?
no

Install Redis extension for PHP?
yes

Install Redis?
yes

Thought I would provide more context. After uninstalling and reinstalling redis, and verifying every site on the server is using the currect LSCache plugin and current redis browser cache plugin, The server stayed up for almost 4 hours and then went back to going up and down.

For the moment I’ve setup a cron to restart LSWS once an hour, until I can find a solution or I move to a different panel.

Exact same issue here, have been constantly tweaking with the settings. No avail. Not sure what’s happening here, whether it’s a cyberpanel issue or a litespeed issue.

So after doing some research this isn’t by Cyberpanel but by Litespeed. Try downgrading your version.

/usr/local/lsws/admin/misc/lsup.sh -b -e 1.7.13

Thanks for the post @mememan. I’ll test it out and report back. Do you have any details on what’s wrong with litespeed so we can watch the changelogs?

Just to confirm, after 12+ hours of rolling back light speed, the server seems more stable. I’m betting this is related to the order of operations in the Cyberpanel upgrade. Because my initial issues started after an upgrade to Cyberpanel. At the time the current Litespeed was only 1.7.13 & 1.7.14 hadn’t been released yet. But obviously re-installing 1.7.13 over what was there was helpful.

Thanks mememan … I rolled OLS back as well to 1,7,13, and it’s gotten better, however OLS is still crashing once per day instead of every few hours. I am wondering if we need to roll OLS back further before 1.7.13 … Cyberpanel 2.0.1 never went down until I upgraded to Cyberpanel v 2.1

How far should we roll OLS back to prevent any crashing?

@pagenet I’ve been 100% stable since rolling back. Did you setup a cronjob to reboot OLS often to avoid the issue? I had one setup but removed it.

The only other thing I did was move one website with a permanent redirect off of my main OLS server(it was always the first to go down). Its just an SSL placeholder for a redirect to Facebook. If by chance you have one of those maybe we found a common thread…

I think my problem has something to do with ModSecurity … all my websites start becoming unresponsive after about 24hrs … restarting LSWS (OLS) does nothing to help.

However, as soon as I disable both ModSecurity rules packs (OWASP ModSecurity Core Rules & OWASP ModSecurity Core Rules), all websites become responsive again without having to restart.

When I re-enable both sets of ModSecurity rules packs, the problem does not immediately return … but probably will in the next 24 hours or so.

Still not sure specifically what the actual culprit is.

@pagenet, Did you ever find a solution that works with ModSecurity?

I am running into a similar issue as explained on this thread and the only thing that works is turned off ModSecurity.

Try to find out the file causing this issue and get rid of that rule file.

@usmannasir I’ve added my comment on the original thread rather than jumping in-between different threads. I appreciate we continue there. Thanks.