Bukkit

@joeyderijk

I'm assuming what you want is a new player spawn where they respawn forever until they do some action that earns them a permission, at which point you want the default strategy to fire? (or whatever you want to put there)

events:
  onJoin:
    - spawnNamedSpawn:noobs

  permission:
    entry1:
      permissions: [no.longer.a.noob]
      onJoin:
        - default

In this example, I've setup onJoin to ALWAYS spawn at spawn "noobs" (which you can set by typing /setspawn noobs). This means new players will always be sent to this spawn on login, even if they logout and log back in 5 times.

Then, I override this global default with a per-permission setting that says if they have the no.longer.a.noob permission, I want to just do run the default rule, which of course will just let Bukkit or some other plugin do it's thing. You can obviously name the permission whatever you like, you just have to be sure that every player who isn't a noob has it.

@violation84

Yep, just as soon as Bukkit releases a beta of 1.6. Until then, you can find dev snapshots for HSP on Jenkins.

Personally I think both are stable enough to have releases by now, but they like to take their time. And until they release a beta, I can't lock in on a version in the maven repositories (meaning anything can change; they proved this conclusively in 1.4.5-R3 when they intentionally broke CraftBukkit naming conventions: http://goo.gl/bFH0W), so I keep it dev until I have a release to lock on to.

@alfun5432

Yes, please read the FAQ. This question is covered there along with a solution.

@ThisUsernameIsMine

HSP self-monitors it's own strategies and events for performance, to make sure if it ever thinks there's a performance issue, you can see it and respond to it.

In this case, it's saying the players named home command took 0.6 seconds to run, which is an unusually long time; it basically means your entire server paused for a half second when the player ran that command.

Now the question is: why? The most likely culprit for this is HSP's safeTeleport mechanism or more specifically, Bukkit chunk loading as a result. If you disable safeTeleport, HSP wil simply grab a location and pass it to Bukkit, which says "send the player here". Bukkit takes that and goes on with life and if the chunks are slow to load, it happens in the background and that one player gets a poor teleport experience, but the rest of the server carries on. So turning off safeTeleport should make the warning (and the brief pauses they are warning about) go away.

With HSP's safeTeleport enabled, HSP is first checking the target blocks to make sure they are safe before issuing the teleport command to Bukkit. In order to do this, Bukkit is forced to load the chunks NOW instead of asynchronously later, so it can respond to HSP's queries about the blocks.

Unfortunately, there is no way around this. If you want safeTeleport on, Bukkit has to load the chunks synchronously in order to respond to block id checks for HSP. If you turn off safeTeleport, you won't have the long strategies, but you also won't get safeTeleport behavior, either. Note Bukkit is not dumb about this, it is only loading the one chunk in question (or two if it's right on a chunk border and the target location isn't safe, so HSP has to look at nearby blocks), so basically Bukkit is taking an unusually long time to load one chunk, which likely boils down to server performance (rotational disk latency + high I/O contention results in long delays when trying to read the region/chunk off of disk).

About the only thing you can do, if you want safeTeleport enabled without occasional half second pauses (as appears to be the case in your server environment), is to get faster disk. Dedicated servers running on SSD should never see this issue since they should be able to load chunks on-demand fast enough to respond to safeTeleport queries and still return in far less than 250ms. The server I used to administer only had traditional hard disks (HDD) but because it was a dedicated server and the disks were never more than 20% busy, I never saw this warning because Bukkit is able to load up any needed chunks very quickly from an idle disk drive. The server also has 16G RAM, so undoubtedly the OS smart cache of disk sectors in memory helped a lot too. I suspect these warnings are a lot more common in a virtualized environment where disk I/O is usually the biggest bottleneck.

It's also possible you're doing something with random strategies; this is the exact same effect except exacerbated by the fact that a random location is LESS likely to be safe then a static home, so HSP may have to search up to 50 blocks away, which can result in several chunks being loaded synchronously in order to find a safe location.

@ttftw42

Just in case you don't see ticket updates, I responded to this issue on the ticket you posted into (thanks for searching existing tickets, it's appreciated). Please read and respond there and/or feel free to open a new ticket if you believe it is a new issue not addressed there.

@limer5

Please open a ticket for #1 and include the full stack trace being printed out into your server.log. Also useful information is your versions (Bukkit server, HSP and other plugins in use) and your HSP config.yml.

My guess without having seen any of that, is the most likely cause of this (since ALL events are failing) is a mismatched HSP/CraftBukkit version. Because of a change the Bukkit team made to the way the core packages are named, you MUST have the correct version matched. The latest 1.7.7-SNAPSHOT available on Jenkins is built for CraftBukit v1.6.2 dev release. If that's not what you're running, go over to the Files tab and download the HSP version that matches the CraftBukkit version you are using.

Issue #2 sounds like maybe HSP is not working (possibly broken in your environment as #1 might indicate?) and that the default Bukkit behavior is being used instead. I can't explain why it would work when they stay on the same world, but let's focus on solving #1 first and probably #2 is related and will go away when #1 is fixed.

@limer5

Yes. It does so out of the box, should be able to just install and go. Make sure your users have the hsp.home.bedsethome permission, though.

And check out the options for bedhomes if you want to tweak how they work.

@MCZq

I'm working with the host provider to get Jenkins back online. In the meantime, feel free to use this Dropbox link for the lastest HSP 1.7.7-SNAPSHOT for CraftBukkit 1.6.2:

https://dl.dropboxusercontent.com/u/1743272/Minecraft/HomeSpawnPlus/HomeSpawnPlus-1.7.7-SNAPSHOT.jar

@SiFlux

Yes, thanks for the headsup, I also noticed it is down today. Server host is cheap and uptime hasn't been 100%, but I'd say it's up 99% of the time. You happen to be trying to get to it during the 1% it's down, so just wait a bit and I'm sure it will come back.

Regarding git/maven/building, it's not a problem most of the time since A) the site is up most of the time and B) dependencies don't change that often so if you've already got a build, you're fine. If you're just trying to get started building HSP, then I'd say you picked an unfortunate time to start, so be patient and I'm sure you'll be back in business soon. Also, all the dependencies I host for maven are simply for convenience, you can still find them all elsewhere if you're desperate.

If you have free hosting to offer with better than 99% uptime, I'm all ears.