MCMA Instance not starting up after crash

  • 360 Views
  • Last Post 24 November 2017
Prohulp posted this 20 November 2017

OS Name/Version: Ubuntu 16.04.3 LTS (GNU/Linux 4.10.0-37-generic x86_64)

Product Name/Version: AMP Instance Manager v1.5.0.0 built 02/11/2017 13:55

Problem Description:

Last night everything worked fine, this morning i woke up and the AMP instance had crashed, so i tried to start it back up with "./ampinstmgr -s MCMA" then it gave me:

[Info] Waiting for AMP instance to start...

[Error] Failed to start AMP. Exit code was 0

[Info] This instance failed to start. Run './ampinstmgr LastLog MCMA' to view it's most recent log.

Ive already noticed something funny:

[Debug] Arguments: -f /dev/null new-session -s "AMPMCMA" -d "cd '/home/AMP/.ampdata/instances/MCMA' --; /home/AMP/.ampdata/instances/MCMA/AMPLinuxx8664 "

Log: https://pastebin.com/1xShrA7G

Steps to reproduce:

To couse the bug to happen, unknown To get these errors, try to start ANY instance

Actions taken to resolve so far:

Tried reinstalling the ampinstmgr, tried using the Nightly version, restarting server, total clean install of AMP.

None worked

Order By: Standard | Newest | Votes
Mike posted this 20 November 2017

Check the most recent log for that instance. The dev null is deliberate to stop tmux from loading its configuration from somewhere else.

Prohulp posted this 20 November 2017

Gladly ive copied the log the first time i saw it, now it shows me this when i try to look into the log:

[Warning] Couldn't start the editor to view the log. Specified cast is not valid.

Prohulp posted this 20 November 2017

The most recent log is attached in the first post ;)

Mike posted this 20 November 2017

Upgrade to Nightly

Prohulp posted this 20 November 2017

Dear mike, i have already tried this, it gave me the exact same error.

Actions taken to resolve so far:

Tried reinstalling the ampinstmgr, tried using the Nightly version, restarting server, total clean install of AMP.

Mike posted this 20 November 2017

The log you've posted shows a release build rather than nightly. Can you produce a new log showing the nightly version please?

Prohulp posted this 20 November 2017

This is correct, the log is from when i was running release build, when i tried running Nightly build, and tried looking into the log it gave me an error:

[Warning] Couldn't start the editor to view the log. Specified cast is not valid.

thats all it showed me, nothing more

Prohulp posted this 20 November 2017

After some looking around on the server ive found the log, im not sure why it doesnt want to display but this is whats in the file:

https://pastebin.com/r74b6frH

Mike posted this 20 November 2017

Try starting that instance manually by running its AMP_Linux_x86_64 executable from its directory - see what happens then.

Prohulp posted this 21 November 2017

I get the same error, nothing diffrent

Mike posted this 21 November 2017

What error? The last log you posted appears incomplete.

Can you do ./AMP_Linux_x86_64 | tee out.txt

And then upload out.txt to pastebin/similar?

Prohulp posted this 21 November 2017

Here you go: https://pastebin.com/00VRhqc2

Prohulp posted this 21 November 2017

I saw in the config file that the module line is empty, wich is wierd since ive never edited the file myself.

Ive added "MinecraftModule" and seems to accept this, now its saying:

[13:45:04] [Core:System Error]    : AMP may not run without a licence.

wich is also weird since this instance worked just fine and hasnt been edited since the crash

Prohulp posted this 21 November 2017

After i have added the module to the config file, it now seems to accept the "./ampinstmgr -s MCMA" command, still errors but this time its showing the "[13:45:04] [Core:System Error] : AMP may not run without a licence." error instead of nothing obvious

Full log: https://pastebin.com/vRFeNbH3

Prohulp posted this 21 November 2017

Since i suspected that the server somehow corrupted its config file, and then generated a new one Which was being created blank made it so that the module and licence was missing.

I created a new instance and this seems to start / work perfectly, for testing sake ive tried removing the config file while running, so it would recreate the same error (i also did this while the server was down, since its unclear if the server was offline)

both cases a blank file got generated, and the exact same errors appear as my first post.

Prohulp posted this 21 November 2017

End conclusion:

It seems that the server somehow managed to delete / corrupt the AMPConfig.conf file, and generated a blank one on restart by command.

It would be nice that the config file dissapeared, it asked the user what module this is, and or restore a backup copy of the file.

This all happened without any information without any errors randomly according to my own logs about 1AM when not a single person was on the server, it was just being idle, no scheduled tasks or nothing

Mike posted this 24 November 2017

I'll look at having the instance manager perform a sanity check on the configuration before trying to start it to capture situations like this and give you a friendlier way of getting the instance back into a working state.

Close