Hi,
I tried to run the Simulator with my monitoring application and a rather large amount of 10k nodes. After a long run, my simulation finishes and log file reads
WARN [SimulationThread – 86]# java.lang.UNIXProcess@1bd15f8c: 2014-01-16 09:25:22,691 ERROR [Simulator – 216]# Simulation finished with the error = null
WARN [SimulationThread – 86]# java.lang.UNIXProcess@1bd15f8c: 2014-01-16 09:25:22,692 INFO [Simulator – 221]# Realtime Duration of experiment (m:s) 1291:12
WARN [SimulationThread – 86]# java.lang.UNIXProcess@1bd15f8c: 2014-01-16 09:25:22,692 INFO [Simulator – 223]# Simulated time is 3:24:37:136 (H:m:s:ms)
Simulated time should be 04:30h, so the run ended prematurely. The error however states null, which is not very helpful. I already modified the Simulator.java and changed the block
try {
scheduler.start();
finishedWithoutError = true;
} catch (RuntimeException e) {
to
try {
scheduler.start();
finishedWithoutError = true;
} catch (Exception e) {
to maybe catch an Exception which not belongs to the RuntimeException class. So far no success. However, reducing the runtime or node count drastically (1000 nodes or just 60mins runtime) seems to work.
My application sends a lot of messages through the network and therefore generates a lot of operations. Is there a limitation towards the amount of generated operations? Or can the reason for simulation failure be something else? Did you ever experience something similar during long runs of the simulator?
Kind regards,
Philipp
.
