RBNBToKepler actors need to timeout more quickly
RBNBToKepler actors need to timeout more quickly when an RBNB server cannot be reached. It currently takes too long for this to happen, and blocks Kepler in the process.
#1 Updated by Derik Barseghian about 11 years ago
Partially addressed by r7986, timeouts now 50% shorter (previously we were actually waiting through 2 OpenRBNBConnection timeouts).
OpenRBNBConnection can still take a patience-trying long time when pointed at a server that isn't running an rbnb server (e.g. ~1m 15s during some tests), but there is probably a good reason for this (dataturbines on flaky connections).