Talk:Server/Requirements/Dedicated

This info might no longer be relevant
With the performance gain switching from Spigot/Bukkit to PaperMC, the scaling of servers specs should be much more lenient compared to a few years ago. There is also performance gain from using different structured server-grade CPUs, newer CPUs with lower clock speed are expected to beat older CPUs with high clock speeds. With great respect for those who wrote this originally this page isn't very helpful right now for the two reasons I mentioned. You will have to do the research yourself or ask on active communities like r/admincraft. As people returning for Minecraft 15.1 I think we should at least point that out. -Gordonhch (talk) 19:50, 7 January 2020 (UTC)

Illogical differences between "Server Grade" and Lesser Servers
The RAM and other requirements for "Server Grade" servers is stated to be higher than for lesser servers. This is illogical... The only reason to run "Server Grade" hardware and OS is that it is supposed to be more efficient. I suspect there is essentially no difference in the RAM requirements to achieve smooth play between a fairly good desktop and a fairly good server. But if there were any difference at all, the server would require less resources to achieve smooth play... That's the entire point of having an expensive server.. In short, all of this is ridiculously moot unless standards (I.e. minimum frame rate or max number of visible loading incidences per minute while running, etc.) are defined and then tested against. Humanjhawkins (talk) 03:02, 10 September 2019 (UTC)

Win95
My cousin's computer has windows 95 and it "runs" MC, why you hate? --Khlorghaal 04:51, 10 May 2013 (UTC)


 * There's no hate, it said "theoretically" due to no support for Java 1.5 on Win95. If it works, it works.  23:52, 11 May 2013 (UTC)

"Recommended" Hardware for Unix/Console not true.
I am hardly able to play with 7 people with a Pentium 4 @ 3Ghz and 2GB DDR RAM in a LAN enviroment. The Server runs MineOS CRUX and eats up ALL the RAM. Symptoms are Connection Errors which kick players forcefully out of the game.

So.. whats up with that. Who even tested these settings?? --213.47.142.126 01:15, 26 May 2013 (UTC)


 * I do not know, but I have a suspicion...If you have an older console, or older parts, that may be the problem. -Creepergoboom64 01:23, 26 May 2013 (UTC)


 * Hm what do you mean by older parts? I set this up last month and updated everything, the os, the scripts,.. btw all of the players are connected through WLAN, does this effect the RAM eating in any way maybe..? --84.114.186.108 23:16, 27 May 2013 (UTC)


 * Unless everything's running on 802.11n hardware, wireless bandwidth is likely to be an issue. The server should have a wired Ethernet connection to your wireless router, and preferably a few of the players as well. I think MineOS intentionally allocates all available memory to the server, since it doesn't expect to be running anything else at the same time. -- Orthotope talk 23:42, 27 May 2013 (UTC)


 * We are running 802.11n here, but maybe not all of the clients support it, havent looked that up yet. the server also has a wired connection to the router. currently i'm testing with 2 players in total which are connected with a cable, played roughly an hour and we are up to 850MB RAM usage while the load is 1:00 — 0.32; 5:00 — 0.22; 15:00 — 0.18; --213.129.230.10 14:09, 28 May 2013 (UTC)

I have to ask who even thinks this is real
Really, have anyone tested these requirements at all? They are all - aside from the "Net Speed" bits - entirely unrealistic. If they're tested at all, what were the settings? Only count ticks once an hour, and a Drawdistance of 1? I'm running my (dedicated, *nix with console only) server on a 3.3GHz core2duo and 4Gb of RAM (3.6Gb dedicated solely to the server), and it starts chugging as soon as a third player enters (and it's on a 100mbit/100mbit WAN-line, all wired gigabit ethernet locally, of course) - and even if only two are around if they're far away from one-another; TOP shows a CPU-load of 1.0 and a fully allocated RAM-pool if both players are far away and working/doing things, or as soon as a third player enters, and this is with vanilla MC without a single mod or any other form of buggery. I'm currently stresstesting it to see how much drawdistance affects it (it's currently at 20)... I could report more detailed testing with my exact server-conf and all, but the nub of the matter is that these requirements are either bonkers and entirely made-up like microsofts "privacy" statement, or they're only true on a server conf written entirely in binary - i.e., all changeable values are either a 1 or a 0. 176.10.248.200 11:58, 21 July 2016 (UTC)
 * Works for me. 1GB Linux MAT-E is a bit slow but works fine most of the time. | AndrewAB (talkAndrewAB.png 12:00, 21 July 2016 (UTC)
 * @AndrewAB - what are your exact specs and server conf, mate? 176.10.140.234 17:46, 21 July 2016 (UTC)


 * 1GB RAM - Linux MATE (Pronounced MATAE) - 4.8 (x86_64-linux-gnu) - (Ubuntu 2015-12-02) - 3.13.0-77-generic Intel(R) Pentium(R) 4 CPU 3.00GHz 2992.501 MHz - Does it really matter -_- | AndrewAB (talkAndrewAB.png 19:15, 21 July 2016 (UTC)
 * Na da, or "Of course" - and well, that is a nice answer, but it only answers half of the questions; ok, so that's your CPU, I can guess that it's using DDR1-RAM at 400Hz, I guess... However, what about the server conf? in particular, the literal file server.conf, and in there, in turn, in particular the Drawdistance, and, well, most of it, really... But if you don't really know what that is, or how to edit it, I know what the default settings are (which you would be running if you've never mucked about with it, e.g. the snooper is on, the Drawdistance is 10, etcetera) - and how many players do you actually - simultaneously - have online? And how much of the RAM i allocated to the server?
 * This is important because the article might actually lie/be false, and in it's current state, I suspect that it is, unless you're running with very low settings (resource-consuming-wise), in which case that should be reflected in the article, e.g. "These recommendations only apply if you're running with a Drawdistance of max 10, and doesn't handle larger loads such as e.g. automated pigman-grinders or heavy redstone-machinery very well" or something along those lines, as I heavily suspect that if you're playing with 2 or more players, that CPU of yours is always maxed out/having a total load of 1.0 or more, and your ram frequently running out. Cheers, 176.10.140.234 20:46, 21 July 2016 (UTC)


 * I play with 1 player, and lowest settings (I call it fine because I'm use to it). This page is obviously been tested out by computers with better tech. Maybe this page really does need redoing. Like testing THEN inputing test results. Also I have a lot of space. | AndrewAB (talkAndrewAB.png 21:13, 21 July 2016 (UTC)


 * Wait, what, mate? This is the "Dedicated Server Requirements" page, you do realize that, right? :-P ... It sounds like you're talking singleplayer - in which case, using the lowest settings, yeah, your rig will work (but then it's also dependent on GPU and so on), but... This is, again, the "Dedicated Server Requirements" page, it has nothing to do with playing solo in your own world - did you mix the talk-pages up? :-P Cheers, 176.10.248.196 23:34, 21 July 2016 (UTC)


 * I said I play WITH 1 player. Me and 1 more. AndrewAB + quelqu'un (AndrewAB + someone). ;) | AndrewAB (talkAndrewAB.png 07:52, 22 July 2016 (UTC)


 * Ah, right, my bad. And yes, that's sort of my gripe here, you can't just guesstimate things like these, you need to actually test first, and then write it down. And hm, if you're playing with one more guy on that hardware, even on minimum settings (which I think should be added to the page, i.e. "These requirements are based on default (low) settings for the server"), it seems that it'd run so-so; have you ever actually observed the server (e.g. load of, almost more importantly, console output); with you and your friend playing, does it never miss a tick, or get an error (many smaller errors, such as the main thread not being able to keep up and skipping a hundred ticks or so, isn't always noticeable to the players, depending on what they're doing) - does that hardware always keep up with you to, or does it often resort to skipping ticks? Cheers, 176.10.140.234 09:48, 24 July 2016 (UTC)


 * It resort to skipping ticks. I think. | AndrewAB (talkAndrewAB.png 11:21, 24 July 2016 (UTC)


 * Tried with Me+2 players and normal settings. A bit slow. Version 1.10 | AndrewAB (talkAndrewAB.png 13:01, 26 July 2016 (UTC)


 * Ah, right - then it isn't really "stable" in the technical sense, if it needs to skip ticks; even if it's not always apparent to players, skipping ticks still mean that the game skips updates to the game-world etcetera. And ah, yeah, I'd assume that, and I appreciate that you actually tested it, well done. But yeah, so with that in mind, the article should probably be updated. 94.254.51.246 10:52, 27 July 2016 (UTC)


 * . Needs an update. | AndrewAB (talkAndrewAB.png 11:20, 27 July 2016 (UTC)