|
Post by Polaris Seltzeris on May 21, 2020 1:11:19 GMT
Just because it's common doesn't mean it's a healthy sign. The issue is not that we need to restart when it's necessary, the issue is that it is ever necessary to restart and no commands can be run. "Fix the server's stability", as has been stated for years now. Without limiting op’s permissions or pulling patches out of thin air, most exploits cannot be avoided when it is tried out on the server by some op or when it’s the fault of minecraft. While we do restart, the prominent cause is a person who would later be banned. When an exploit or way to abuse the server is found, we try and patch it. It's hard to come up with solutions for exploits when there's no platform of discussing them. Even Seth said on another thread that he can't fix exploits if nobody has explained them to him. The idea that we have to limit freedom to fix exploits is hogwash and even if it's a Minecraft issue it's something that can be solved with some basic server sanity checking.
|
|
square
Veteran Member
Asst. Creative Designer
Posts: 1,294
| Likes: 1,291
|
Post by square on May 21, 2020 1:16:15 GMT
Without limiting op’s permissions or pulling patches out of thin air, most exploits cannot be avoided when it is tried out on the server by some op or when it’s the fault of minecraft. While we do restart, the prominent cause is a person who would later be banned. When an exploit or way to abuse the server is found, we try and patch it. It's hard to come up with solutions for exploits when there's no platform of discussing them. Even Seth said on another thread that he can't fix exploits if nobody has explained them to him. The idea that we have to limit freedom to fix exploits is hogwash and even if it's a Minecraft issue it's something that can be solved with some basic server sanity checking. No, some recent exploits cannot be fixed without limiting freedom and the prevention of crashing via methods would do the same. We cannot currently stop the restarts we have each day (restarts which only take 20 seconds) as these are either newly made when we have not fixed them or just using minecraft against us. I would like to know how you would plan on ‘fixing the server’s stability’ however.
|
|
97
Veteran Member
RIP Telnet
Posts: 889
| Likes: 679
|
Post by 97 on May 21, 2020 1:20:46 GMT
Quick update, now the server is starting to stop without any context and always has to be tf!killed.
|
|
|
Post by Polaris Seltzeris on May 21, 2020 1:59:22 GMT
It's hard to come up with solutions for exploits when there's no platform of discussing them. Even Seth said on another thread that he can't fix exploits if nobody has explained them to him. The idea that we have to limit freedom to fix exploits is hogwash and even if it's a Minecraft issue it's something that can be solved with some basic server sanity checking. No, some recent exploits cannot be fixed without limiting freedom and the prevention of crashing via methods would do the same. We cannot currently stop the restarts we have each day (restarts which only take 20 seconds) as these are either newly made when we have not fixed them or just using minecraft against us. I would like to know how you would plan on ‘fixing the server’s stability’ however. Again, it's hard to come up with solutions for exploits when you don't know what they are. "Some recent exploits cannot be fixed without limiting freedom" and "prevention of crashing via methods would do the same" are extraordinary yet vague claims. If you'd like to know my plan then you have the ability to read through all these stability related threads in which I provide ideas, and then such ideas are ignored by everybody, and here we are again discussing the same exact things.
|
|
|
Post by Polaris Seltzeris on May 21, 2020 2:00:18 GMT
Quick update, now the server is starting to stop without any context and always has to be tf!killed. What does "starting to stop without any context" mean? The daemon just suddenly terminates?
|
|
square
Veteran Member
Asst. Creative Designer
Posts: 1,294
| Likes: 1,291
|
Post by square on May 21, 2020 2:22:16 GMT
No, some recent exploits cannot be fixed without limiting freedom and the prevention of crashing via methods would do the same. We cannot currently stop the restarts we have each day (restarts which only take 20 seconds) as these are either newly made when we have not fixed them or just using minecraft against us. I would like to know how you would plan on ‘fixing the server’s stability’ however. Again, it's hard to come up with solutions for exploits when you don't know what they are. "Some recent exploits cannot be fixed without limiting freedom" and "prevention of crashing via methods would do the same" are extraordinary yet vague claims. If you'd like to know my plan then you have the ability to read through all these stability related threads in which I provide ideas, and then such ideas are ignored by everybody, and here we are again discussing the same exact things. You’re repeating what I said about how we cannot fix exploits that are in the future, the prominent cause for crashes or freezes are some exploit or other necessary measures that are out of our hands. I doubt that a stability fix will iron these out. While the case for this is here - its rare and I’m sure other admins or players can vouch about the reasons the server halts. edit: and those claims are true and I could list some recent ones now
|
|
|
Post by Polaris Seltzeris on May 21, 2020 2:27:47 GMT
Again, it's hard to come up with solutions for exploits when you don't know what they are. "Some recent exploits cannot be fixed without limiting freedom" and "prevention of crashing via methods would do the same" are extraordinary yet vague claims. If you'd like to know my plan then you have the ability to read through all these stability related threads in which I provide ideas, and then such ideas are ignored by everybody, and here we are again discussing the same exact things. You’re repeating what I said about how we cannot fix exploits that are in the future, the prominent cause for crashes or freezes are some exploit or other necessary measures that are out of our hands. I doubt that a stability fix will iron these out. However, in the case of this thread it is rare - I’m sure other admins or players can vouch for my former sentence anyways. How are they out of our hands? If it's a server exploit or a crash, you can fix it in the code with developer effort. Changing our infrastructure, such as to have our server use multiple daemons instead of one, would offer us more stability so that it can't crash or freeze up the server as easily. Clearly our single daemon is being overloaded constantly if it's having to restart multiple times every day...
|
|
square
Veteran Member
Asst. Creative Designer
Posts: 1,294
| Likes: 1,291
|
Post by square on May 21, 2020 2:36:10 GMT
You’re repeating what I said about how we cannot fix exploits that are in the future, the prominent cause for crashes or freezes are some exploit or other necessary measures that are out of our hands. I doubt that a stability fix will iron these out. However, in the case of this thread it is rare - I’m sure other admins or players can vouch for my former sentence anyways. How are they out of our hands? If it's a server exploit or a crash, you can fix it in the code with developer effort. Changing our infrastructure, such as to have our server use multiple daemons instead of one, would offer us more stability so that it can't crash or freeze up the server as easily. Clearly our single daemon is being overloaded constantly if it's having to restart multiple times every day... I can agree with you on the daemon part however some things out of our hands are just lag machines or recent exploits. These are out of our hands as (as both of us said) it’s impossible to care for recent or future problems (like exploits) instantly including crash chests and arrows. When the exploit is found it’s (usually) patched and forgotten about and I’m just stating that the majority of freezes and restarts (which you touched on) is due to human intervention and could not possibly be fixed - quote me if I’m wrong.
|
|
97
Veteran Member
RIP Telnet
Posts: 889
| Likes: 679
|
Post by 97 on May 21, 2020 2:38:45 GMT
What does "starting to stop without any context" mean? The daemon just suddenly terminates? Yes
|
|
|
Post by Polaris Seltzeris on May 21, 2020 2:53:22 GMT
How are they out of our hands? If it's a server exploit or a crash, you can fix it in the code with developer effort. Changing our infrastructure, such as to have our server use multiple daemons instead of one, would offer us more stability so that it can't crash or freeze up the server as easily. Clearly our single daemon is being overloaded constantly if it's having to restart multiple times every day... I can agree with you on the daemon part however some things out of our hands are just lag machines or recent exploits. These are out of our hands as (as both of us said) it’s impossible to care for recent or future problems (like exploits) instantly including crash chests and arrows. When the exploit is found it’s (usually) patched and forgotten about and I’m just stating that the majority of freezes and restarts (which you touched on) is due to human intervention and could not possibly be fixed - quote me if I’m wrong. The solution to lag machines is to 1. make it easier for admins to easily locate and monitor them 2. look into resetting/removing/updating chunks with lag machines automatically (if a chunk is taking way too long to process) 3. change our infrastructure so that a lag machine can't just significantly slow down the whole server, we do too many things on one daemon. Crash chests and arrows sound like problems that can be fixed with sanity checks, but since I have no in-depth explanation of what even happens with those, I can't even theoretically come up with a solution nor can anybody else without information. I posted a suggestion a while ago which was ignored which would basically set up a system which splits our daemon into a cluster, each individual daemon would manage a service such as an in-game world, or some plugins, and there would be a system which connects each cluster which allows the game experience to be the same for an end-user as if it was just one.
|
|
|
Post by Polaris Seltzeris on May 21, 2020 2:53:46 GMT
What does "starting to stop without any context" mean? The daemon just suddenly terminates? Yes That can certainly be debugged with Java then, if it literally just exits...
|
|