kvimber / MCStation

Minecraft Server Software
8 stars 1 forks source link

Server Jar Not Finishing Startup Process #1

Closed kvimber closed 13 years ago

kvimber commented 13 years ago

This is for the issue where the server jar file stops finishing the startup process.

You (Kendrick) said that you saw this on Windows. I tried to reproduce it last night, and couldn't on Linux.

Test on Windows and if the problem is there, we'll take a look further into what's going down.

Current Problem:

I've setup the MCServer.run_test_setup method that should kick everything off, and start up the server. It should make reproducing this problem easy, as it waits for a full four seconds after startup to check for logs (and then dumps all logs since last startup) so you should be able to easily see if the system gets setup all the way. You can just comment out the run_server method call if you'd like to be able to debug further, or if four seconds isn't enough to generate a world on your system.

ghost commented 13 years ago

I ran the run_test_setup with an environment containing no server existing. The server builds to 85% before getting stuck. If I run the java command outside of python it builds the entire server to a startup point.

kvimber commented 13 years ago

Well if you use run_test_setup and you see it get to 85% in the log printout that comes with that method, then that just means that it got that far in about 4 seconds. If you just start up the system, and then keep pumping out logs using this command: mc.format_logs(mc.get_logs_since_last_start()) you'll be able to see where the logs are until there. So if you pump the logs for like a minute, you never see it pass 85%?

ghost commented 13 years ago

Yeah I will open the actual log text file and keep refreshing it with no change. I'm unable to connect on the server side as well.

kvimber commented 13 years ago

alright, I'll look up this problem when I get home to my windows computer tomorrow.

ghost commented 13 years ago

An update on this - I updated the java memory arguments to 1536 on each and that seemed to resolve it.

kvimber commented 13 years ago

tested this on windows and it was working for me, so I'll close this now that we're both not seeing an issue.