After reboot and tracking cpu usage using top firewalld uses around 20% CPU but it increases exponentially until it freezes the whole server after sometime after reaching 100%.
The repeated ‘already enabled’ makes me think firewalld and fail2ban are out of sync. Long story but firewalld must start before fail2ban. If firewalld got restarted that has been a problem in the past. There is a fix in the firewalld config to keep the fail2ban stuff during restart but it is now disabled by default. I think people restart the firewall wanting a clean/fresh restart. This isn’t good when you have other programs relying on it though.
EDIT : Also, a large swathe of 138.199.x.x appears in that screen shot. That could be a large scale DDOS? As a quick test you might add 138.199.0.0/16 as a temp block.
Thank you for your help, i’ve done both, resynced fail2ban and firewalld and added that ip range as a temp block to see if it works. I’ll keep watching and update with the results.
EDIT : Also, a large swathe of 138.199.x.x appears in that screen shot. That could be a large scale DDOS? As a quick test you might add 138.199.0.0/16 as a temp block.
Blocking the IP Range solved the cpu usage issue, thank you!
I suspect that means fail2ban is messing up the rules. It shouldn’t be seeing traffic from the same IP once a rule has been created.
They show up once for port smtp, 465 and submission, but when blocked they dont show again, i think this 3 ports are the same arent they?
The reason I said to do a temp ban is that’s one hell of a lot of addresses to block. Seems to be Amsterdam for the main block but these get sold and traded so hard to say. Seems I’ve seen a fair number from the Netherlands though considering how small a country it is.
But, depending on the server’s purpose, you may see no impact on your services other than less spam.