Eric J Ehlers
Comments
The experience describe with this comment (https://bugs.mojang.com/browse/MCPE-21416?focusedCommentId=1093385&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-1093385) ) could possibly be this bug. I have a few others on that thread that might be as well
It's getting very difficult to refrain from obscenities about you Minecraft. There are definitely going to be daily posts about the lack of progress on the bug.
NEW EVENT YAY! (Why in the #### I'm even still playing is a mystery, this is so screwed up)
Switch, single player, local.
Need a metric ton of nylium. So planning to go to the Nether. I hates the Nether, so I do the closest to cheating I do on my survival world, I make a copy of the world.
Go into the copy, make some progress in the Nether, I hit a mile stone, make a second copy. Now, it appears that Switch has a 1GB limit on saves. As the world is around 300MB, that limits me to 3 copies of this world. I now have
original world
The first copy (same original state, now with progress)
The second copy (same state as first copy's progress)
I load second copy, play a bit and die because of some glorious screw up. I weight the options. I have a bunch of mats (and my elytra) gathered in an ender chest, so could just rebuild armor/gear and go back, or restart at the milestone. As rebuilding the gear will take something like 40ish diamonds and 100ish enchantment levels, I decide it is LESS bother to gather the mats again. I attempt to make a copy, forgetting I already have 3 copies. The copy fails (after like an hour of waiting (hyperbole!) through the hang at 40% it always does) Grumbling, I delete The first copy, bc it's a dup of second copy. And make third copy.
If this were all logic, this leaves an identical set of copies as previously, but with one of them having a different save date/time
I enter third copy, get all the rest of the mats I need (including an accidentally discovered ancient debris! WHEEEE!)
Return to overworld. Hey? What happened here? There are no doggies at my base's cave entrance. Where are Huey, Louie, and The Third One? And where are the 2 as yet unnamed but obediently sitting cats? I venture inside. Where are my four horses with saddles and one with diamond armor? (sadly, they are not Famine, Death, War, and Pestilence, bc that is what I named my pigs). And Shadrach, Meshach, and Abednego the llamas are gone? As is Riker the donkey and the unnamed but beloved mule I bred and raised from a child? GURGI IS MISSING! MY CAT GURGI IS GONE!
Suspicious, I venture to the livestock pen. My 16 multicolored sheep (names will be given on Twitter, but I'll spare you here), the cows Jessie, Bessie, Tessie, and Frank, the pigs (named above) and half a dozen chickens (all from thrown eggs) are not here? It looks lonely.
I check the bees. The bees, (at least some of them) are still there. Whew. That's a bigger pain to replace. Welp, the village is over 100 blocks away, surely they will be untouched. Nope, 32 good men and true gone, all named and traded heavily with, and all de-zombified at least once. As are the 5 named cats. And,.. uh.... the named and penned zombies I have to keep them in line and highly discounted (Thelma and Louise, if you must know).
There was only basic lag that comes from travelling on an elytra with fireworks through a dense area like the Nether. No crashes. Departure to the Nether was within 100 blocks of the missing animals (but, as said, not so close to the village). Egress is further away due to multiple Nether portals close together in the Overworld (original gate was next to Roanoke).
I close and save copy 3, and try to decide what's a bigger burden, regathering the materials or rebuilding all the mobs? Well, lets see where we are. Check Copy 1 (progress after first copy of the day), which is in the nether. It takes a few minutes to get out, but I dont' worry about preserving much bc I just want to know. Nope. All gone.
Check original world. Same deal. WTF Mojang. Either a normal copy deletes mobs, or else a failed copy deletes mobs from the world it is copying and the world that was copied from. This is beyond tolerable. The fact that you spout "Minecraft is a game people can trust" while this stuff is happening is offensive. Get somebody to communicate what you're doing NOW.
Excellent observation. I used an Elytra for the first time yesterday. Experienced this. It's in Switch. 1.17.whatever is current (too lazy to open it up). Why is this marked as resolved?
New event that appears different from the majority of my losses.
mandatory on Switch, play session was about an hour before this happened. Single player, local.
I took a horse on a longish journey. 1100-1200 blocks west maybe? died. Respawned near point of origin. Took a second horse to same location. Penned both horses, did my bidness, tamed a couple cats, then headed back riding Horse #2, with Horse #1 on a lead. When I reached Roanoke, I put Horse #1 in a mine cart bc I have a track to my base/spawn point. Pushed him then followed on Horse #1, both cats still following. This is a stretch of ~300-400 blocks south east. More south, but the majority of the track is diagonal.
When I arrived at base, horse #1 and minecart were at the stop point. I went to the stable to deposit horse #2, a move of approximately 10-20 blocks in both X and Y. Went back up to get horse #1, but both it and the minecart were gone. At this point I noticed only one cat was still with me. The start of the mine track has a bed, so I thought it might be there. And I thought maybe the minecart got bumped when I moved the other horse, so I went back to the start of the track. No cart, horse, or cat.
Addendum: I played about another 1.5 hours, mostly messing with my trading hall. I needed another villager so I went to check on my breeders, 100 blocks from the trading hall. Both adults (named farmers, but not traded with) were missing. There was a child villager, however. There is a 9x9 area on which the farmers can move, though I can't guarantee it's within a single block. The child villager entered the stream which I use to transport them automatically to the trading hall, but never arrived at the other end. I went through the stream, and he wasn't there. It is conceivable he could have drowned, but, there's a constant flow. To drown the AI would have to have made him stay in a single block of water source that has a full block of air below it (buttons to keep it out. If he drowned, he was the first in well over 30 villagers to do so. I reckoned he de-spawned like his dear old dads. (They are named Adam and Steve). The breeders were located a good 100 blocks from the spawn point as well.
Extraneous: I'm fascinated by @Tarjei Aasen's report. The Y coord grouping is bizarre. It doesn't match all my experiences (the horse that despawned was at Y63 and the villagers at Y~40, so today's does. But my Roanoke Catastrophe took place at and above Y65. Well, that's not accurate. 6 of the cats were something like, 59, 53, and 47 in the basements. But all the livestock, other cats, birds, dogs, and villagers were all at (at least) Y65. So, assuming the same bug, Y64 might not be anything special, but Y-space grouping is interesting. Isn't a chunk from bedrock to top of the world? Or am I remembering that wrong?
@David Robinson. How does it feel to give me hope then crush it?! 😃
I tested my village. I dug down to Y 11 in multiple places in Roanoke. Only one of them turned up any open space, and it was mostly water. I think it unlikely that this was the fate of my villagers. Also, Some of my lost animals had open levels I had built below them when they despawned, but they weren't there.
However, due to the fact that chunk restriction seems to help in some cases (but only some) it would seem that there are likely multiple bugs causing this. If the chunk load is responsible for some, and your explanation is responsible for some others, my guess is there are 3+ causes. Which... seems to be a case for refactoring the entire save process, imo. If it were my project and I had this degree of failure with multiple causes, I would be starting from scrap. Probably reuse a lot of functions, but I can't imagine just putting a band aid on it wouldn't cause more bugs and less readable code. It'd be like the kitchen of an Italian restaurant in that repository.
@Goldenhelmet – there does seem to be a misconception that chunk borders are "the thing" – perhaps because the "workaround" is posted so prominently. But with so many cases of immobile entities (horses in stalls, villagers in a trade hall, etc) – this seems to be a misleading workaround. It's clear many of the users are focusing on that specific cause, though. Perhaps either demphasizing the 2020 workaround, or at least more strongly wording the info in the workaround that it is not intended to be a fix for all situations, and only may help some situations. Though, to be honest, that is explicitly stated in bold text already, so... people just gonna read what they want to see, I guess.
I have had a few minor losses recently that involve lag (world is now > 290 MB, the Switch takes some time to render when it's saving) but not crashes (I've not had Minecraft crash or close without a full save in quite a while). Nothing on the scale of "roanoke" that I mentioned above, but a pig or a cow here and there. Most recently it was a pig and a cow – both babies – while the other four cows and four pigs were unaffected. All 10 animals were in the same pen which probably crossed chunks. The disappearance happened after a 3ish hour play session. I did close/save and when I rejoined the next day the babies were gone. (only listing mobs because the baby - parent connection may, somehow, be connected?). I have lost child villagers one or 2 at a time this way in the recent past as well.
Yay! New experience! Several incidents since my last comment, but this one was weirder than most. Sorry for the length, but I'm a story teller, and I promise I tried to keep it only to facts directly relevant.
Set up: I have three bases in a single player, local world.
1 next to my original spawn point. This is currently where all my animals are kept. At the time of this incident, probably 2 dozen chickens, 40 or 50 rabbits, 3 horses (with armor and saddles), 4 llamas, 3 donkeys (with saddles and chests), a dozen cats (all named and "sitting" so stationary), 2 dogs (stationary), 3 parrots (named and stationary). Immediately next to the base is a village I constructed that houses 21 villagers, all of whom have been named and traded with to at least level 2. The village is surrounded by a fence, so they can't move too far off and only spiders can get in (extremely rare) plus generally 2 golems spawned automatically (they seem to despawn and respawn with startling regularity. Also some crops, bc convenient.
2 is where I'm working my big project ~200 blocks away on z axis, maybe 300 on x axis. It's a lot of construction, and even more materials. Other than a minecart and 8 or so chickens created by throwing eggs, no entities that I expect to be there from time to time, just a disturbing number of llamas abandoned after their wandering trader was either killed by mobs or (more likely) despawned.
3 is a minor base camp roughly 2000 blocks z axis away from the big project. It has a single entity, a dog. This camp exists bc his world was created before Caves and Cliffs release, which means every place I had visited previous to the release would have no copper (or other new materials). It's where I go to gather copper or amethyst.
I have a current need for a lot of copper, so I was visiting the basecamp frequently over many play sessions of perhaps 2 hours average length, I did not track time, but there were several visits, each time starting a new strip mine tunnel (one direction, so a few hundred blocks at a stretch) with enough time strip mining to fill up a couple shulker boxes (I mine EVERY ore out of habit), then return to main project (2) to smelt the ore and sort all the minerals and drops found. Most of these visits include a short visit to the village (1).
During my last three visits I made an exception to the new tunnel and kept on the same one. In addition to exploring a mine I bumped into, the tunnel ended up being maybe 2000 blocks long, all going further away from all three bases. Nothing unusual after the first 2 trips, but after the third trip, I do my usual smelting and sorting at base project (2). I made a building using maybe 120 blocks (pretty small, considering one of my structures at this base is literally an artificial mountain). Then a visit to the village (1).
At this point the village is EMPTY of villagers. In addition nearly all my livestock, including the rabbits, AND all the indoor pets are just... gone. Literally, there are now something like 4 chickens and a sheep at this villages.
MOST curiously, however, I come to this now empty village and for the first time ever, CATS have spawned. At least 3 cats (they are not my cats, as they do not have names or collars and run from me). I thought that cats spawned based on the population of the village. I reckoned previously they hadn't spawned due to the relative proximity of my tamed cats, but that's just a guess.
Also probably relevant: It has been suggest that a combination of distance and time from the village/base (1) is the cause for this. I find this unlikely. The first village I cultivated, a few hundred blocks further from my mining (and on the far side of the village from it, making the lost village, now provisionally named Roanoke) still has it's entirely population, which is also around 20 villagers, a few cats, and a golem.
There were no system crashes, though I did, at least a few times, save and quit moments after going into the world because I discovered I needed to leave the screen for several minutes and was not in a safe spot.
This was all single player, local play. travel distance is discussed above. Player spawn point changes frequently, as each visit to a base usually includes a visit to a bed, mostly to not have to mess with phantoms. I cannot recall any times where I placed a bed, created a spawn point, and then picked up the bed again. There was most certainly lag at times, bc the world is 250mb and the switch takes a long time to do it's excessively frequent saves, wherein rendering often doesn't happen at all. This lag is always during game load or during fast travel (not something that happens when mining). Travel between bases takes less than a minute due to my system of mine tracks and nether portals.
TL;DR or for those who want a victim count, I lost 21 villagers, 3 armored horses, 2 back donkeys, 4 llamas, a couple dozen chickens, several dozen rabbits, 2 pigs, 2 cows, 3 birds, 2 dogs, and 12 cats in one fell swoop. More distant villages unaffected. Village population based spawn behavior still observed at empty village.
Note from @unknown: the cats spawned because the game still considers missing villagers to be part of the village population until their 20 minute timer runs out. The missing tamed cats probably had been blocking cat spawns before the snap.
The job block connection is interesting. I'm not sure it would be helpful fixing the bug, but it does seem to indicate that not all the data is removed. The entity... still exists somewhere.
Reply from @unknown:
The game stores village data separate from entities. Villages check regularly to see if their dwellers are still present. When a villager goes missing the village keeps checking for it for 20 minutes before releasing the link between that villager’s ID and its workstation, bed, and bell. This prevents immediate loss of links and breeding for an empty bed if a villager is outside the village or in unloaded chunks for a brief period of time. The same applies to cats and iron golems, too. After the 20 minute timer expires the village will also check the villager’s last known position if it is in an unloaded chunk, before removing the villager from the dwellers list.
What does "ADO" mean? there's a severe lack of good documentation about the bug process and field meanings, esp if Bedrock is working differently than java, et al.
Reply from @unknown: documentation can be found here.
So here's a different use case than I knew before. Previous mob despawns seemed to have no connection to any sort of crash. This one is only... almost a crash, I guess?
Anyway, was copying a world on my Switch. World copy failed, repeatedly, at 40%. When I logged back in, 3 named and levelled villagers were gone (already replaced with new child villagers by the time I saw it, though it wasn't more than a few minutes.
This bug literally is the reason I won't spend any money on Minecraft anymore. My kids want a family realm, but this is game breaking. And it really really sucks that Mojang refuses make it a "Mojang Priority."
Continues to be an issue, for the record.
Then you need to talk to your PR team, Dr. Awesome. Bc deleting comments looks bad. Look, I'm a developer, I know it's cluttering up the ticket. But Mojang is otherwise silent on the issue, and no mods or devs have said anything about progress for literal months. Fans want to know they are heard, but it doesn't feel like Mojang cares in any way about what even you admit is game breaking.
I know bugs where you can't be certain how to repeat are the most vicious and hard to do anything about, but when devs say nothing and the company says nothing, users are left only to feel that you guys don't care.
I suppose this comment will be deleted too, but I just needed to let you know that this problem is bigger than an in-game bug. You have a PR bug, and the only fix for that is to make sure users know Mojang/MS are putting effort into fixing the code bug.
In attempt to save this comment, let me point out something that you could be using the chatter for, even from a bug ticket perspective.
1) Reproducing this is clearly not a simple {{if A and B then C happens immediately }}— because hyou don't know what A and B are. You don't even know if it's just A and B or if it's A through E.
You're going to have to reproduce actual play conditions. Which is not clean like most bugs. And this means looking into comments when people say, essentially, "me too" – if that's ALL they say, ask for more detail. Your team should be hunting for the details, not casually waiting to see if you get lucky and a user reports something easy. If they posted, they care enough to comment, which means they may be willing to provide more if they are prompted.The only thing that makes a bug unrepeatable is not knowing the conditions.
I'll volunteer what I remember of my situation.
I built an enclosed glass structure 18w x 18l x 10h blocks (interior measure). it was filled with water to a depth of 4 blocks, and at the time of the first incident, it contained three turtles and 4 axolotls – all seven were named. The was a small amount of land at the top of the water, so both mob types could leave the water if the AI so chose. The entire enclosure was located inside a warm ocean biome and was completely submerged.
A few days later, 2 drowned spawned inside the enclosure. When I showed up, one axolotl was playing dead, and another was missing. I assumed it had been killed by the drowned. I killed the drowned and put more lighting in to prevent hostile mob spawning.
The next time I left and came back, the previously wounded axolotl was missing as well.
The next day I caught another axolotl and named it. Then I bred 3 more (which remained unnamed). At this point I had six axolotls that I had interacted with by breeding or naming.
Two days after that, all six axolotls disappeared. Apparently at once. I had been mining and shaping the immediate environment (removing dead coral, planting live coral, mostly). None of the turtles have ever gone missing.
Note that this isn't a quick repeat. You might have to let devs or testers play for a while to see the bug, and log every single thing that happens in that time. It will mean slogging through tons and tons of data. But frankly, you have a 4 year old bug that breaks the game and infuriates the users. Your current approach is insufficient.
1b) Note that developers or mods asking for more details shows the users that someone is actually paying attention to them. It will help your PR problem at the same time.
2) You could also tell marketing, business, and customer relations about every single report. I know that at my job for a national bank with thousands of customers, if we report every time a customer complains (esp if they're all complaining about the same thing) then the problem becomes a priority pretty quick. Resources get allocated, and fires get lit under people so they put more attention into the issue instead of doing the "insufficient data, not my problem" thing.
As it is though, all the users have is this one Jira ticket, which has almost no mod or dev interaction, and the bulk of that interaction is deleting comments. You're alienating your user base which reduces the game popularity. As I said before (the comment was deleted), the action I'm getting closer to taking is not buying the Java Edition for my new computer. It's dropping Minecraft all together.
So I spent some time playing with this in creative mode last night. I'll upload some pictures if you feel they are necessary.
I started experimenting with a 2(w) x 9(l) x 4(h) block open space (interior measure) surrounded by glass (but no roof) so I could see the inside. I put water along the entire top, then to make sure it was all source blocks, I grew kelp up through it all, then cut the kelp down. This is how I've done bubble elevators in the past, so I'm confident it works.
Then I placed 4 fish in the top level of water (It's all with tropical fish, so assume that when ever I talk of fish)
The fish didn't block each other, but they did go to the bottom and stay there. They did (rarely) move up to the second block, though. So I added a fifth block to the height. The fish are still disposed to the bottom most block, but would occasionally rise to block 3.
My conclusion is that one behavior of tropical fish, at least if the player places them, is that they avoid the top two blocks of depth in a given volume of water.
Wondering if crowding could be an issue, I added 1 fish at a time, pausing to observe behavior between each, until I reached 18 fish total (the number of water source blocks on each level. The behavior remained unchanged. There were some blockages, but it was more like a traffic jam than being stuck in place. The fish could get unblocked and often did.
One observation is that it is possible for more than one to occupy a block , at least if they are moving. However, if they have a head-on "collision" they have a bit more difficulty getting out of each other's way. They won't enter a block with another fish in it if they're both not moving. Or maybe if only one or they other, but they start to get jammed when this happens. If there's too many, they get clustered up doing this.
But there's something else very weird going on.
My original plan was to make a room where the walls were fish a big fish tank. So my next step was to change this wall structure and make it a square with a hollow center. Each side of the square was 2 (w) x 9 (l) x 5 (h) interior measure. Using the same method as above, I made sure each block was a water source. Then I broke the glass blocks separating the original wall from the new part of the structure. I did not add more fish.
The fish spread around the whole square, but began to get jammed next to the corners. Note, it wasn't in the 2x2 space shared by joining walls, it was right next to these spaces. They fell into 3 unequal clusters and stopped being able to free themselves.
I don't know what about the shape of the container changes this, but it's weird.
Weirder still is when I blocked off the corners with glass blocks to return one of the sides back into a 2x9x5 volume – the fish did not unblock each other. Even if I moved through them to scatter them, they went right back. At this point, I removed a block from the bottom, and the fish tried to go into the newly opened space.
I killed several fish to see if that would unblock the rest, but it did not. Once jammed, they stay jammed.
Side note that feels related: in a survival world I have made a big glass cube in a warm ocean biome where there is 2 blocks of water between the roof and the surface. Tropical fish and octopuses get stuck here a lot, and there's always a half dozen swimming forward at the top edge of the top glass block.
As I said, I've got screenshots of all of this and a little bit of video, but it's on my Switch, so kind of a pain to share. If it will help, I'll make the effort though.
Updated.
Yes it does. I am running a few more tests. I think water behaves really weird, even for Minecraft Physics (tm). But I've got more of a hang of it and want to figure out some more.
The most recent update did not correct this behavior. But playing around gave a bit more data. Moving my character into the squares occupied by the fish disrupts them, and in a small enough space they may move to a different elevation. However, they eventually move back to the lowest point
Your comment ignores the fact that Mojang, in the ONLY almost-communication they have done about this has assigned THREE ADOs to this issue. Meaning they consider it three different issues. So no, we don't all know it's the same problem.