Martin Bittermann
- cubetrace
- cubetrace
- Europe/Stockholm
- Yes
- No
MC-86513
![[Mojang] Grum (Erik Broes)](../../../data/assets/gravatar/grum.png){kind=small}
Martin Bittermann: Thanks, I split them up and made you the reporter of both tickets since the original reporters are both inactive.
Martin BittermannMartin Bittermann: Thanks, I split them up and made you the reporter of both tickets since the original reporters are both inactive.
I can confirm this issue on 1.8.4 SMP server, no mods.
Will try to make a video when it happens again.
Have been able to film a siege in our village:
https://youtu.be/zSi_QxOCwf4
This shows clearly that in some situations, golems simply do not care about some zombies.
I assume the status should be changed to ReOpened.
Not a duplicate of
MC-858.I can confirm this bug on vanilla 1.8.x servers. This happens often to me when mining blocks with instant-speed in Survival. It's a server-client synchronization issue and has little to nothing to do with lag.
To make the distinction clear:
MC-858describes the internal server having low TPS and therefore registering the block-breaking action from the client very late. This leads to the client re-displaying the block again, until it gets the confirmation from the server that its actually broken, seconds later, or not at all. At the end, the block is either broken or still there, on both client and server.This bug (
MC-5694) is different in that it can occur on a server running 20 TPS constantly. It only occurs when the time to break a block is instant, e.g. with a diamond pickaxe on Nether rack or high-Efficiency diamond shovel on dirt. While the block in question disappears from the client's view (it is not rendered anymore and is removed from clientside collision), the server says it's still there and pushes the players' movement back when the client tries to move within the block's space. This can be a quite tricky and dangerous situation, especially when taking down a pillar below you: You fall into the ghost block again and again (many times a second) and cannot move. Client and server arrive at no consensus over whether the block is still there or not. Reloading the chunks on the client by pressing F3+A does not resolve this. It can be resolved by trying to place a new block in the old location. This of course fails server-side, the new block is not placed, but the old one reappears client-side. It can also be resolved by rejoining the server.So this is a persistent client-server asynchronicity and must be fixed independently from the (in my opinion) non-bug
MC-858.By the way,
MC-7409is a duplicate of this.Can confirm this. We had a lot of hopper pairs facing into each other across chunk borders on or SMP server, in an attempt to create chunk loaders. We had a single item of cobblestone oscillating in each hopper pair. Upon inspection after a few weeks, most hoppers had 2-4 stacks of cobblestone. I first thought those had been left there on accident, when digging the tunnels, but turns out this was the item duplication bug. Then I recreated the setup in our base, with 3 pairs of hoppers, two of them at a chunk border, with different items, one item per pair. All three had about 4-7 items in each hopper, overnight.
I'm not sure whether it has something to do with the server messages "[Server thread/WARN]: Fetching addPacket for removed entity". I had some of these, too, but only 3 that night. One other player was logged into the server at that time.
Did anyone notice that the comments here describe two completely different phenomenons?
MC-5694andMC-7409.I hope the mods can untangle this somehow.