Rudder

In fact the documentation (which states it is a connection timeout) is wrong, it is applied as SO_RCVTIMEO on the socket, and hence closes the connection when we see no output for 30 seconds, which may happen in changes only mode for example.

Interestingly, it's the agent that closes the connection.
Agent side

   error: SSL_read: receive timeout
rudder  verbose: Connection to 192.168.42.3 is closed

then server side

rudder  verbose: 192.168.42.2> MD5=918083b6dad780baf4e93367aa1cfe7f: Client is TRUSTED, public key MATCHES stored one.
rudder  verbose: 192.168.42.2>      Received:    EXEC 
rudder     info: 192.168.42.2> cf-serverd executing cfruncommand: /opt/rudder/bin/rudder agent run -uR
   error: 192.168.42.2> SSL_write: underlying network error (Broken pipe)
rudder     info: 192.168.42.2> Sending failed, aborting EXEC (send: Broken pipe)
   error: 192.168.42.2> SSL_write: (-1 SSL_ERROR_SSL) bad write retry 
rudder  verbose: 192.168.42.2> Unable to reply with terminator. (send: Success)
rudder  verbose: 192.168.42.2> Remote peer terminated TLS session (SSL_read)
rudder     info: 192.168.42.2> Closing connection, terminating thread

I configured the timeout everywhere

PR https://github.com/Normation/rudder-agent/pull/146

Applied in changeset rudder-agent|faf1c876786b97ac6bf070988cda7f1671c92e61.

This bug has been fixed in Rudder 4.1.10 and 4.2.4 which were released today.

• 4.1.10: Announce Changelog
• 4.2.4: Announce Changelog
• Download: https://www.rudder-project.org/site/get-rudder/downloads/