When "Keep remotely disconnected tabs open" is enabled, once the connection has been closed by the remote host, Royal TSX starts doing port checks in 5-second intervals to indicate the remote service availability. This results in multiple "Did not receive identification string from <my IP address>" errors being logged in /var/log/secure which would in turn trigger a DDoS protection system, and effectively cause an IP ban. For example, a commonly used fail2ban service has a ssh-ddos configuration that does exactly that.
Since there is currently no alternative safer way of checking the SSH service's online status, please make the connection state monitor separately configurable, so it can be disabled while keeping the valuable "Keep remotely disconnected tabs open" feature on.
Perfect! Thx for letting me know!
According to my test it works as expected, thank you very much!
a new beta which includes this feature has just been released.
Please give it a try and let me know if it works as expected!
we implemented the setting to disable connection status monitoring for disconnected tabs.
I don't think your other suggestions would make for good usability. Probing the connection only when a tab is focused means that the displayed status might not be up to date as a tabs can be visible but not focused (like when using external windows). Also just having the status stuck at "Available/Green Light" after it came online again might be misleading as the connection might drop before the actual service listening on the specified port has gone down. So this might lead to false positives or simply showing incorrect status as well.
I think the setting to completely disable this feature makes sense and so that's what's been implemented now and will be included in the next release.
Additionally, even when monitoring is enabled, it would be good to not repeat probes for ad hoc connections, only send probes for currently focused tab, and not send probes when connection is alive. The latter suggestion comes from my thinking that the reason for the feature is for the user to know when the server comes back online. I'm just thinking that limiting amount of probes will limit probability of being mistaken for a DDoS attack.