Something closing SSH access on Plesk Onxy [on hold]

I am running a new VPS server from Heart Internet running Ubuntu 14.04.5 LTS‬ and Plesk Onyx 17.0.17. I have enabled the Firewall extension and have modified the ssh port to use something non-standard. This port is allowed through the firewall, and after applying the rules, I have been able to connect just fine with my SSH key (passwords disabled). When I come to connect a day later, I get a connection timeout (port 22 still gives me connection refused).

The Heart Tech support won’t fix this problem as it’s an unmanaged server (although this same setup worked on an old VPS I had from them), but did suggest that it could have been some sort of cron job and pointed me at this Plesk 12 forum post. I’ve not been able to fix my setup.

I don’t think this is an issue with Plesk specifically (off-topic alert), but more another script or something resetting IP tables or sshd. I’m looking for assistance to track down the problem, although a fix would be fantastic.

I have poked around various log files but don’t quote know where to start.

UPDATE: It’s not Plesk related, but more my firewall settings. SSH service is correctly listening:

sshd    23650 root    3u  IPv4 7376014      0t0  TCP *:XXX (LISTEN)
sshd    23650 root    4u  IPv6 7376016      0t0  TCP *:XXX (LISTEN)

So I am trying to understand the iptables -L command to figure out what is causing the changes. I can see before I re-apply the rules there is no reference to my custom port, but afterwards I see the line:

ACCEPT     tcp  --  anywhere             anywhere             tcp dpt:XXX

I just need to track down what is changing this and work backwards from there. I did ask this on the Plesk forum, but their support chap said it wasn’t a Plesk issue but more of a general server admin task. I don’t want to be stuck between SO and Plesk forums trying to solve this.


Leave a Reply