Occasionally (e.g., every few weeks), the GTFS-realtimeGenerator seems to stall and stop producing updated GTFS-realtime data.
We configured this project to write log to file as explained in CUTR-at-USF/onebusaway-application-modules#17, and this is the log that we captured when this issue happened:
...
2013-04-21 06:04:09,324 INFO [RetrieveTransitDataV2.java:60] : Executing SELECT query...
2013-04-21 06:04:09,355 INFO [HartToGtfsRealtimeServiceV2.java:166] : [901369, 902701]
2013-04-21 06:04:09,371 INFO [HartToGtfsRealtimeServiceV2.java:336] : tripUpdates = 2
2013-04-21 06:04:09,371 INFO [HartToGtfsRealtimeServiceV2.java:337] : vehiclePositions = 2
2013-04-21 06:04:13,355 INFO [Slf4jLog.java:67] : Stopped [email protected]:8088
Unable to access jarfile cutr-gtfs-realtime-hart-1.0.0.jar
"Stopped [email protected]:8088" seems to be the Jetty server shutting down, but I'm not sure why.
Doesn’t seem to have any other abnormal info in log other than “Unable to access jarfile cutr-gtfs-realtime-hart-1.0.0.jar” – and the only references I can find to this on Google are prior to starting the application, not during application execution. It could be triggered by an error in HART’s SQL Server, but I don’t have any evidence to verify this. From the log, and from randomly checking HART database, it seems to happen when HART database is still up – although, it could be caused by an intermittent connection issue that is fixed by the time I test again after getting this message.
It's possible that this could be caused by memory issues, since it seemed to start after we configured the application to output to a log file. These log files have been growing to 500MB or more.
Recently, we stopped logging output because of this concern, so we will see if this issue happens again.