I'm encountering a rather unusual problem when using SSH to connect to one of my routers. My connection remains stable for precisely two minutes before I'm abruptly disconnected with the error message "client_loop: send disconnect: broken pipe". Subsequently, any attempt to reconnect immediately results in a "connection refused" error, which lasts for exactly two minutes. After this period, I'm able to reconnect, only for the cycle to repeat itself.
I've looked into potential solutions and tried adjusting the TCP keepalive settings, as I came across suggestions that this might help with connections being dropped during periods of inactivity. However, this doesn't seem to be the root of my problem since these disconnections occur even while I'm actively using the connection.
I dont have any indication that dropbear isnt running smoothly, so i assume so.
I previously had some logs about maximum concurrnet dns queries and i tried disabling dns to resolve it, but it only got rid of the dns query logs. The ssh isssue did not get affected and still had the same pattern of 2 min on 2 min off.
I could be wrong, but this does not seem to be related to DNS at all. This is either some device or firewall in the path blocking the traffic, or the dropbear daemon failing down.
Right, thought it was something else that i missed. In that case, yes, they all agree on time. I have multiple routers all running the exact same configuration, but for some reason 2 of them have the issue i explained in the initial post.
Yes, it answers if the router are remote and their order. The OP said one router but now it is two; which is no problem just making sure it always was 2 and not spreading.
Which two is it?
Also, is there a physical way around the jumphost?