Jump to content

Recommended Posts

Posted

First, I know have searched around a lot for this error and usually it is because the mod is poorly writen or the entity is being spawned on the wrong thread. As far as I know, either of those are happening here. I am creating and spawning the firework entity on the server side, not client side. The ConcurrentModification is seemly random, but only occurs when there is no block above the player (firework spawned at player's location).

 

Here is the crash log. It is for the server. The clients get read time outs for the server crashing.

 

And here is the relevant code. Along with the "helper" function to create the firework. I got the code to create the ItemStack for the firework from net.minecraft.ItemFirework (I think, it was a while ago, cannot remember). It generates a Firework with identical NBT tags as one Minecraft generates (tested with NBT explorer).

 

I am not sure if this is a Forge bug or me doing something wrong. Either way, some help would be great.

Thanks.

 

EDIT: One last thing, I know at about 150% certainty that THIS code is causing THAT exception. We have actually having this error for almost a Week now and we have tried everything we know to fix it. This is the most possible simplified version of the code that still causes the crash.

Posted

Hi

 

I gather this error doesn't happen very often?  This exception occurs when the collection is modified (eg elements added or removed) while you are iterating through it.  So either the client thread is writing to this collection (trackedEntities), or you have done something in EntityTrackerEntry.tryStartWachingThis() that alters the trackedEntities.

 

Are you really, really sure that you're spawning the entity on the server side?

 

I'm not sure what the object 'this' is in your helper function, but perhaps .side is not correct?

If you put a breakpoint in your helper function just before the spawning, you could check the thread to make 100% sure it is the server thread.

 

Alternatively you might try putting a try {} catch {} block around the EntityTracker.func_72788_a(), and put a breakpoint in the catch.  If you are lucky, you might notice a pattern in what the client thread is doing at the time of the exception.

 

If the error occurs fairly often, you could add breakpoints in your helper, and also in the java.util.HashMap methods with modCount++  (eg  .put(K,V) ) - when the helper breakpoint occurs, enable the HashMap breakpoints and resume.

 

If it's fairly rare, you could perhaps create your own copy of HashSet but containing println methods to log all calls to HashMap methods which affect modCount, for example

 

    public boolean add(E e) {

        System.out.println("MyHashSet.add modCount" + map.modCount);

        return map.put(e, PRESENT)==null;

    }

 

Change trackedEntities to be MyHashSet instead of HashSet and with any luck the log will tell you exactly who the culprit is.

 

-TGG

Posted

It is important to know where the code that you call "relevant code" comes from. What is

this.player

? What is

this.side

? Where does that "relevant code" get called? You can only ever ever call it from the main minecraft thread.

 

this.player and this.side are exactly what you might think. They are initialized when this class is (when the player logs in). This class is for a player's "skill" (RPG-ish mod). So, this.player is the owning player of the skill (EntityPlayer). It is generated upon login for a player. Likewise, this.side is returned from "FMLCommonHandler.instance().getEffectiveSide()" and initialized when the class is; so this should only ever be running on the server.

 

Also, it does only run in the main Minecraft thread (it did not use to, removing it from its own thread greatly decreased the crash amount, but did not stop it). 

 

EDIT: That crash log is a server crash log. The clients get read time out errors.

Posted

Hi

 

this.player and this.side are exactly what you might think. They are initialized when this class is (when the player logs in). This class is for a player's "skill" (RPG-ish mod). So, this.player is the owning player of the skill (EntityPlayer). It is generated upon login for a player. Likewise, this.side is returned from "FMLCommonHandler.instance().getEffectiveSide()" and initialized when the class is; so this should only ever be running on the server.

 

Also, it does only run in the main Minecraft thread (it did not use to, removing it from its own thread greatly decreased the crash amount, but did not stop it). 

 

EDIT: That crash log is a server crash log. The clients get read time out errors.

 

When I put a breakpoint in addEntityToTracker, I see two relevant threads:

"Minecraft Main Thread"

and

"Server thread"

 

Your helper function must run in the Server Thread, not the "Main minecraft thread".  Perhaps you should try putting a breakpoint in there to make sure.

 

-TGG

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Unfortunately, your content contains terms that we do not allow. Please edit your content to remove the highlighted words below.
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Announcements



  • Recently Browsing

    • No registered users viewing this page.
  • Posts

    • Ah, it appears I spoke too soon, I still need a little help here. I now have the forceloading working reliably.  However, I've realized it's not always the first tick that loads the entity.  I've seen it take anywhere from 2-20ish to actually go through, in which time my debugging has revealed that the chunk is loaded, but during which time calling  serverLevelIn.getEntity(uuidIn) returns a null result.  I suspect this has to do with queuing and how entities are loaded into the game.  While not optimal, it's acceptable, and I don't think there's a whole ton I can do to avoid it. However, my concern is that occasionally teleporting an entity in this manner causes a lag spike.  It's not every time and gives the appearance of being correlated with when other chunks are loading in.  It's also not typically a long spike, but can last a second or two, which is less than ideal.  The gist of how I'm summoning is here (although I've omitted some parts that weren't relevant.  The lag occurs before the actual summon so I'm pretty confident it's the loading, and not the actual summon call). ChunkPos chunkPos = new ChunkPos(entityPosIn); if (serverLevelIn.areEntitiesLoaded(chunkPos.toLong())) { boolean isSummoned = // The method I'm using for actual summoning is called here. Apart from a few checks, the bulk of it is shown later on. if (isSummoned) { // Code that runs here just notifies the player of the summon, clears it from the queue, and removes the forceload } } else { // I continue forcing the chunk until the summon succeeds, to make sure it isn't inadvertently cleared ForgeChunkManager.forceChunk(serverLevelIn, MODID, summonPosIn, chunkPos.x, chunkPos.z, true, true); } The summon code itself uses serverLevelIn.getEntity(uuidIn) to retrieve the entity, and moves it as such.  It is then moved thusly: if (entity.isAlive()) { entity.moveTo(posIn.getX(), posIn.getY(), posIn.getZ()); serverLevelIn.playSound(null, entity, SoundEvents.ENDERMAN_TELEPORT, SoundSource.NEUTRAL, 1.0F, 1.0F); return true; } I originally was calling .getEntity() more frequently and didn't have the check for whether or not entities were loaded in place to prevent unnecessary code calls, but even with those safety measures in place, the lag still persists.  Could this just be an issue with 1.18's lack of optimization in certain areas?  Is there anything I can do to mitigate it?  Is there a performance boosting mod I could recommend alongside my own to reduce the chunk loading lag? At the end of the day, it does work, and I'm putting measures in place to prevent players from abusing the system to cause lag (i.e. each player can only have one queued summon at a time-- trying to summon another replaces the first call).  It's also not an unacceptable level of lag, IMO, given the infrequency of such calls, and the fact that I'm providing the option to toggle off the feature if server admins don't want it used.  However, no amount of lag is ideal, so if possible I'd love to find a more elegant solution-- or at least a mod recommendation to help improve it. Thanks!
    • When i start my forge server its on but when i try to join its come a error Internal Exception: java.lang.OutOfMemoryError: Requested array size exceeds VM limit Server infos: Linux Minecraft version 1.20.1 -Xmx11G -Xms8G
    • Also add the latest.log from your logs-folder
    • Add the mods in groups
  • Topics

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.