[Buildbot-devel] "Connection lost in a non-clean fashion" / duplicate slave

[Charles Lepple]

On Nov 17, 2014, at 4:28 AM, Harry Percival <harry at pythonanywhere.com> wrote:

> In the twistd.log, I'm seeing these messages at 
> around the same time:
> 
>     <timestamp> [Broker,<n>,<ip>] duplicate slave <slavename>; delaying 
> new slave (IPv4Address(TCP, '<ip>', <port>)) and pinging old 
> (IPv4Address(TCP, '<port>', <other-port>))
>     <timestamp+10s> [Broker,<n-1>,<ip>] BuildSlave.detached(<slavename>)
> 
> What could be happening here?

If the IP addresses are the same, chances are you have some sort of stateful firewall which is silently dropping the original connection. This might be at odds with the first suggestion at the MariaDB site - a higher Buildbot keepalive timeout would need a correspondingly higher OS-or firewall-level keepalive timeout to not drop the connection.

-- 
Charles Lepple
clepple at gmail