Waterfall - server might have crashed: not accepting accepting console commands or crash-unknown-cmd-message is not set -- restarting.

[TomLewis]

I have dealt with this before; https://github.com/gsand/mark2/issues/45

2021-08-12 17:07:41 | [WARN] Error whilst handling query packet from /51.79.64.238:50914
2021-08-12 17:07:41 | [RAW] io.github.waterfallmc.waterfall.utils.FastException: No Session!
2021-08-12 17:07:56 | [/REMOVED:61192|CaptainProCrazy] -> UpstreamBridge has disconnected
2021-08-12 17:07:56 | [/REMOVED:61192|CaptainProCrazy] <-> DownstreamBridge <-> [survival] has disconnected
2021-08-12 17:07:58 | [WARN] Error whilst handling query packet from /51.79.64.238:49342
2021-08-12 17:07:58 | [RAW] io.github.waterfallmc.waterfall.utils.FastException: No Session!
2021-08-12 17:08:14 | [WARN] Error whilst handling query packet from /51.79.64.238:42639
2021-08-12 17:08:14 | [RAW] io.github.waterfallmc.waterfall.utils.FastException: No Session!
2021-08-12 17:08:23 # server might have crashed: not accepting accepting console commands or crash-unknown-cmd-message is not set -- restarting.
2021-08-12 17:08:23 # stopping piratecraft (caused by crashed)

But this time it's on my bungee and its not related to a an unknown command, BUT there are a lot of [WARN] in my logs for Waterfall, this IP pinging the server is my online user widget on my website, if you put in the IP 51.79.64.238 into a browser ytoull see my website.

Does mark2 restart the server because of too many [WARN] messages? Thinking its crashed?

[Column01]

Are you using the example bungeecord mark2.properties from the samples directory?

If not, you should try it. If you are and its still not working, send a random unknown command and share what bungee says in reply for the unknown command message. This is essentially the same problem as the issue you linked, the unknown command message is wrong for mark2 to check the server is running

[TomLewis]

Yes I am. I failed to mention I have been using Mark2 for well over 6+ years now.

https://gist.github.com/TomLewis/7cc5ab64363ac24b7b95c091fd183c5c

When I run a random command in Waterfall, nothing happens, there is no output telling me that I entered a wrong command. I wonder if that is the cause of a plugin filtering my console output, I know I read something like that in BotSentry.

Let me try disabling that filter and see if it changes anything.

[TomLewis]

Edit: Disabled "consolefiltering" in botsentry and put in some garbage into the console, and still nothing no error message for "wrong command"

[Column01]

did you restart bungee after disabling the console filtering? something is capturing the unknown command message before mark2 can process it

[TomLewis]

Nope, it's a production server.

The plugin allows hot-reloading settings.

[Column01]

I don't think this is a Mark2 issue, this has something to do with how waterfall in conjunction with mark2 is configured. The unknown command message does not make it to mark2 therefore it kills it. If you want to just disable the crash checking to get around this (since a bungeecord server should never crash anyways...) you can add: plugin.monitor.crash-enabled=false to the mark2.properties and use .reload to reload the config for mark2. It should still check for ping timeouts (I.E. Network disconnection issues)

[TomLewis]

I forgot to change the setting for my mornings reboot to see if stopping it filtering messages! What I will do first is try that for tomorrows scheduled reboot and see if the unknown command message works, on top of that I will install waterfall locally and make sure it 100% has an unknown message.

Sadly all bungees are susceptible to crahing with multiple issues with how they handle packets, I have a relatively active server so get targeted alot. If you wanted to test crahing any bungee, anyone that has query open to the public can have their bungee crahsed!

[Column01]

Running the latest waterfall: In this case, it's inside byobu cause we don't use mark2 for our proxy. If your message doesn't show, something is 100% filtering it out.

[TomLewis]

Update: I did actually keep the filtering off last night, must have been tired. yes it was that filtering causing the message to not show up. Now I get

[20:05:50] [main/INFO]: Command not found
[20:05:51] [main/INFO]: Command not found
[20:05:53] [main/INFO]: Command not found
[20:05:54] [main/INFO]: Command not found
[20:05:56] [main/INFO]: Command not found
[20:05:57] [main/INFO]: Command not found

Do you think that was the cause of the crashes?

[Column01]

It most certainly is, if mark2 doesn't find a reply to an unknown command, it assumes the server has crashed and will restart the process. Try enabling it again now that the message is not filtered out.

On Fri, Aug 13, 2021, 3:07 PM Tom Lewis, @.***> wrote:

Update: I did actually keep the filtering off last night, must have been tired. yes it was that filtering causing the message to not show up. Now I get ```[20:05:44] [main/INFO]: Command not found [20:05:50] [main/INFO]: Command not found [20:05:51] [main/INFO]: Command not found [20:05:53] [main/INFO]: Command not found [20:05:54] [main/INFO]: Command not found [20:05:56] [main/INFO]: Command not found [20:05:57] [main/INFO]: Command not found

Do you think that was the cause of the crashes?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/gsand/mark2/issues/134#issuecomment-898664765, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGKMHOJPEL54SNWEG2LZSS3T4VUPTANCNFSM5CBTHF7A . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .

[Column01]

Gonna close this for now, If you need any more assistance regarding setup (I.E. there are no programmatic errors or exceptions), joining the discord server is the best way. Link