OK, Before I start, I really don’t want anyone defending MineOS and what it is designed to do. I have a problem and justifying how MineOS is built does not help resolve this issue. Yes I’m using and I like the fact that its a pure OS that I can run minecraft from.
I DREAD however having to open the WebUI. Why? because it takes upwards of 5 minutes to load my servers. Yes, that long! I Open the IP (its a local ip address), I add my user and password and the page starts to load and then it hangs for a long…loong time before I see my server list. Other times the page will just not load and ill have to shut it down and try again.
So I have no idea what this issue could be. I have 1G RAM allocated to running the PC that is my server (the rest of the RAM is allocated to the games). Do I need to reserve more for this? i don’t know. I would like to have the webUI open like any other website, ie faster than a dial up connection. As I said, I’m not looking for an explanation, I’m looking for a solution and things to try to resolve this. by justifying why its happening is basically telling me to “suck it up, that’s the way it was built” which I refuse to believe.
I have had the same issue since I first started using MineOS (years ago, at this point).
For me, I found that if the worlds had lots of restore points, or archives, they would take far longer to appear in the server list.
Unfortunately, I haven’t found a solution to this, other than manually deleting a bunch of the archives and restore points, which kind of defeats their purpose, I know. I had asked previously about enabling an auto-pruning feature for them. Something like:
For the first day, keep snapshots every hour.
For the first week, keep snapshots every day.
For the first month, keep snapshots every week.
And so on, so that there isn’t as many of them.
I’ve recently upgraded my server quite substantially performance-wise, and the issue has been lessened quite a bit. The UI still takes about 5-10 seconds or so, but it’s down from the 30-90 it could take previously.
For reference, I currently have nine servers, seven of which are online, with maybe 20-30 restore points and a couple of archives each. This is running on an unRAID Docker instance, with 2C/4T of CPU allocated, and however much RAM it wants. I think it’s using about 6GB currently.
From the terminal, check ‘top’. Is the server load high? Is the CPU for the nodejs process the culprit?
Do you use dynmap, or other map-generating services, that generate on the order of tens of thousands or more files?
What commit (the version) is your webui at?
Do you have an excessive amount of restore points?
This problem is pervasive, but I can never solve it because I’m never able to experience it; I just develop MineOS, I don’t play Minecraft, so for the most part, real-world cases where things take long like this…aren’t things I’ve ever had a chance to diagnose and hope to fix. Often, I just shoot in the dark or make changes I know that don’t cause regression but are pretty much still hail mary’s.
1GB of RAM dedicated is more than enough, so also check how much RAM (and report back!) your nodejs process is using. Is it increasing over that 5 minutes?
Really sorry, but I dont understand where to find this ‘Top’ The server usually has around 1 - 1.5G of free RAM (Aside from the 1G used to run the PC) The graph display usually sits in the Yellow shade. it has been sitting with the blue shade at the top, but in thois isntances I have a High amount of RAM Free (up to 4G)
Yes I do, but I have this set up so that it updates slowly. Mind you, Id find it hard to believe this is the main cause of the issue, Ive had Dynmap running on a number of other server hosts and never had an issue like this one.
Guessing this is the version? git commit: 23f10ac
No very few, still trying to work out the cron job feature.
Sorry I can be of more assistance here, but I’m more than happy to work with you and give you as much infor as I can so you are able to make a more informed decision.
@Hexparrot:
If you want (for testing purposes, and to try to root out any bugs using a running server), I’ll give you access to the WebUI and shell of my MineOS server. Just tell me first if you need to reboot. I’ll tell you what servers are the most active (and smart to let be). Other than that you are free to play around.
(Oh, and I edit Line1004 in mineos.js to make the archive name windows filesystem legal; I use “_” between hours, minutes and seconds)
you definitely need to at least load up a test server and instance with dynmap and chunk pregenerator mod and figure out why those screw up mineos beyond belief like they do, i tried to pregenerate my world the other day just doing a 300 chunk radius circle from the spawn point and it is literally impossible to get the server to stop gracefully with either the pregenerator or might only be dynmap, but that’s part of my start up tasks before i can play i need to pregenerate the world as well as complete a dynmap fullrender of it ideally.
They “Screw up” Mineos like they do because they demand a huge amount of resources. The same with a huge amount of mods, and running pregenerationof worlds.
MineOS is dependent on the underlying OS doling out resources as it seems fit. If you run other software on top of it demanding proority and loads of resources, MineOS cannot do a damn thing about it.
MineOS is a nodeJS based script that runs on a lightweight webserver on an as small footprint as possible to be able to use as little resources as possible. MineOS only acts as a WebUI remote control panel to start and stop minecraft servers. What you ask the servers to do in addition is your problem, and MineOS cannot do anything about it.
i have a huge amount of resources available, it should still work just fine, i went out of my way to have 16gb of ram and a decent if late model processor dedicated entirely to minecraft, it works a lot better with other server hosters i’ve seen as well, mineos is the only one which produces issues with dynmap.