Forcing a crash with Cauldron?

[Targren]

I'm trying to deal with an odd problem where our server crashes, but it doesn't "crash-crash"; that is, the process stays alive, but is in a state where it doesn't respond, the server loop doesn't appear to be running, etc... This is particularly obnoxious because it means that neither auto-restarts, nor notifications work to get it back on it's feet (spigot restart-on-crash just makes it spam the server log with "Unknown command" errors forever)

 

Is there a way to induce a crash with Cauldron (git-Cauldron-MCPC-Plus1.6.4-1.965.21.89 (MC: 1.6.4)), the way F3-C crashes the client? I suspect that the problem is in one of the plugins, but without a reliable way to make it crash (instead of just "at the worst possible time" the way it seems to choose on its own), it's going to be tricky to troubleshoot.

Edited August 7, 2014 by Targren

[plowmanplow]

• Try it with the latest version of Cauldron for 1.6.4. Blood hit his paetron goal so is now updating 1.6.4 with a lot of fixes from later versions.
• You might get better luck asking in Cauldron's IRC channel.
• It is possible to pull a stack trace from a running server if you have the JDK instead of the JRE using jps (to get the java process ID) and jstack (to pull the stack trace)

[Targren]

 

• Try it with the latest version of Cauldron for 1.6.4. Blood hit his paetron goal so is now updating 1.6.4 with a lot of fixes from later versions.
• You might get better luck asking in Cauldron's IRC channel.
• It is possible to pull a stack trace from a running server if you have the JDK instead of the JRE using jps (to get the java process ID) and jstack (to pull the stack trace)

 

 

I didn't know he'd hit the goal, so I'll try that first. The IRC channel was no help. No one could seem to understand why I'd want to do that.  

 

Thanks!

Edited August 7, 2014 by Targren

[jominer247]

If your cauldron doesnt want to stop, kill java

 

Multicraft:

Spam the stop button xD

 

Windows:

Task Manager > Stop java/javaw.exe

 

Linux:

Depends on OS, For CentOS

$ pidof java

and then end the PID.

[plowmanplow]

You may have approached the IRC channel from the wrong perspective. You may have better luck going in like so:

• State that you seem to be having a hard-freeze/threadlock issue with your server.
• State that you have a stack trace of the frozen/locked server and provide a pastebin link to the stack trace.
• Wait for relevant response.

[Targren]

 

You may have approached the IRC channel from the wrong perspective. You may have better luck going in like so:

• State that you seem to be having a hard-freeze/threadlock issue with your server.
• State that you have a stack trace of the frozen/locked server and provide a pastebin link to the stack trace.
• Wait for relevant response.

 

 

Wow... I completely blanked on "threadlock." It would have made explaining my problem that much clearer.

 

I think I've been working too hard.

[jominer247]

You can never work to hard when it comes to servers, Make your server good and fun for all the players, and you'll see for yourself how rewarding it is

[Targren]

You can never work to hard when it comes to servers, Make your server good and fun for all the players, and you'll see for yourself how rewarding it is

 

It's not my server, I'm just the volunteer admin: but I was referring to working too hard at actual work (where I also admin non-MC linux servers, as well as other things...)

[jominer247]

You can never work to hard on that either

[Targren]

 

You may have approached the IRC channel from the wrong perspective. You may have better luck going in like so:

• State that you seem to be having a hard-freeze/threadlock issue with your server.
• State that you have a stack trace of the frozen/locked server and provide a pastebin link to the stack trace.
• Wait for relevant response.

 

 

The "Relevant response" I got is the not altogether unreasonable suggestion that I try to reproduce the problem without cauldron(using stock server, I assume). The problem with doing that is that I don't know how to reproduce the crashes, and doing that live and waiting for the crash to come about organically is pretty much impossible: all of our anti-griefing protections come from plugins: it would be a disaster.

 

Now I'm reduced to trying to set up crayfish ovens and crafter pipes in a distant, unused, and deletable region, to "inspire" the server to crash...

[Xiaos]

Try the stop button in Multicraft or end java/javaw.exe in the windows task manager.

[jominer247]

Xiaos. I don't use multicraft, but even if I did clicking the stop button once doesnt help. In order to KILL the server (forcefully shut it off) spam the stop button about 10 times.

[Targren]

Try the stop button in Multicraft or end java/javaw.exe in the windows task manager.

 

I'm  not trying to end the process: that's trivial (/kill <pid> on linux): I'm trying to force a crash. Two very different things.

Edited August 12, 2014 by Targren

[Nentify]

OpenEye has a configuration to crash on start up, and isn't required on the client if you're using a modpack that is not your own.

[Targren]

OpenEye has a configuration to crash on start up, and isn't required on the client if you're using a modpack that is not your own.

 

This. This right here. EXACTLY what I needed. THANK YOU.