How so?
I'm not 100% sure where it falls apart. I'm suspecting the timeouts on both ends of the equation. It takes some time for RLFOSSIL to timeout a
connection, and then there's the timeout HAPROXY end... I'm not sure which
end is really holding it in limbo.
HAPROXY appears to accept connects from the outside, while losing any
available connects on the inside. I suspect HAPROXY has reset itself before RLFOSSIL either gets reset or decides the connection is gone. Is it RL
hanging up? It can be susceptible to this, but either way getting hammered doesn't help it.
At times you can see some of whats happening, have a look at the VM's while some are busy and you're trying a local connect (still goes through HAPROXY) VM's not busy.
Whichever it really is, the port change already seems to have had a large impact on availability.
Spec
*** THE READER V4.50 [freeware]
--- SuperBBS v1.17-3 (Eval)
* Origin: Good Luck and drive offensively! (21:3/101)