They need to rename Gleba to Bartleby

[NineNine]

I am going to explain the "why" of pausing and restarting Gleba Production in depth here. I thought I had provided enough details for others to understand, but apparently, not.

Post Script:

While writing those posts, I was running a new design for a factory on Gleba in the background while not watching it.

I was using the following recipes: Agriculutural Science Packs, bioflux, yumako processing, jellynut processing, pentapod eggs, and nutrients from bioflux; all through bioreactors.

Guess what?

All I found was spoilage backing up and no damage to any part of the factory (turrets to prevent freshly hatched pentapods for bioreactor/science-pack-production from attacking and destroying stuff).

From what I could tell, something interfered with the bioflux-to-nutrient production for long enough, which caused all the bioreactors to run out of nutrients to keep running. But what is the exact sequence of events that led to this interruption in nutrient production? I only have guesses and potentially terrible speculative fixes.

I also tried using an assembler with a worse spoil->nutrient recipe (no nutrient needed to start) as a backup source in an earlier Gleba factory design, but that failed in the past for different reasons.



And you thought I was joking? I'm sorry to disappoint you. No, I'm not even trying to make jokes about my frustrations with Gleba.

I was planning to add a rocket fuel (from bioflux and jellynut) sub-factory to this new design when I spotted the completely dead-still bioreactors on Gleba. Yet I couldn't do anything as there was no bioflux or jellynut to test/work with for rocket fuel production.

When I have had a bad day at work and want to come home to take my shoes off and relax, troubleshooting a Gleba factory is not my idea of a fun and relaxing afternoon.

Again.... you should consider asking for help. Please post in Gameplay Help and we'll help you get Gleba figured out. There's no shame in not figuring the whole thing out all by yourself. Posting in here over and over again how you can't figure it out and it's the most terrible game ever made because you can't get it, and nobody else can figure it out either, isn't going to help you get to Aquilo. Obviously, many thousands of people have already figured it out, and several of us are reading these forums, ready to help.

[aka13]

Eh, I would not say that there is any sort of consensus of course, or "people drop the game because of gleba", that would be absurd/in bad faith, imo as well.
I just find it interesting to observe, that there is indeed an observable difference between the planets.

It is the most difficult(in first run) of the three but in the same time the most interesting and rewarding.

I would say that it is "most different" to what one usually expects from factorio, that is certainly true, I agree with you.
However I did not perceive it as rewarding or interesting, but to each their own. Out of the 200h I had in the first and only run I spent about 30 on Gleba.
"Solved" it in on-demand-ways to produce science and the products, and never went back.

[mmmPI]

I would say that it is "most different" to what one usually expects from factorio, that is certainly true, I agree with you.
However I did not perceive it as rewarding or interesting, but to each their own. Out of the 200h I had in the first and only run I spent about 30 on Gleba.
"Solved" it in on-demand-ways to produce science and the products, and never went back.

You prefer Vulcanus which is the easiest planet by all metric ! not everyone enjoy sit back gaming some of us wants things that are challenging !

[aka13]

No need to throw around empty accusations. There is nothing complicated in spoilage. There is literally is not even a way to do JIT-logistics cross-planetary for them, where the real complexity would be at.
I shut down gleba production completely, and enable it on-demand, that is more complex than the intended solution already. If having a timer on items is complex for you, well, I guess that is nice.

SA is so far a regression in terms of difficulty either way, since dw marathon is not a thing anymore, and planets are still unaffected by presets, if you want to talk about "gamey" complexity.

[mmmPI]

No need to throw around empty accusations. There is nothing complicated in spoilage. There is literally is not even a way to do JIT-logistics cross-planetary for them, where the real complexity would be at.
I shut down gleba production completely, and enable it on-demand, that is more complex than the intended solution already. If having a timer on items is complex for you, well, I guess that is nice.

SA is so far a regression in terms of difficulty either way, since dw marathon is not a thing anymore, and planets are still unaffected by presets, if you want to talk about "gamey" complexity.

It's no accusation ! I thought i remember correctly when you said it to me on discord that Vulcanus was your prefered planet. You even showed your early belt setup and ask if someone had a better one and the several ways of solving gleba were discussed.

I think SA adds a lot of additionnal complexity, with quality more than with spoilage as dealing with spoilage is not optionnal, but also not as difficult as the math involved for finding the most efficient way of making quality things.

[aka13]

Oh, okay then. Yeah, systems-wise SA brings a lot of new fun stuff, totally agreed.
And yeah, youremember correctly, I did perefer vulc, and left gleba for the end. I didn't like the continuos flow stuff, absolutely.

[Jay_Raynor]

When I initially tested some defense strategies on Gleba, I found that very little to nothing about defending against biters/worms was applicable to Gleba natives.

A defense perimeter of turrets and walls? Stomper walks over the wall and potentially becomes aggressive due to the presence of turrets.

Strafer spawns wriggler on top of the turrets and, as the namesake implies, strafe instead of moving in close.

I didn't ask for new enemies to act the same way as Nauvis fauna. I accepted and dealt with Vulcanis Demolishers of various sizes, for instance.

Yet, at the same time, Gleba's fauna is different enough that new players will have problems adjusting after Nauvis.

No, you want them to behave same as Nauvis from that checklist. You don't have complaints about Demolishers because you don't defend your factory from them. Vulcanus is all-offense. Nauvis teaches you what the basic tools are, how to keep them running, managing enemy spawners in the factory area to stay ahead of evolution, placement of defenses for interlocking fire... Do you even landmine? It's okay to be frustrated that you're struggling with a solution to a particular Factorio problem. You used the blueprints of others to facilitate a speedrun, so I'm not sure why you don't just simply look up Gleba defense strategies.

I am not mixing those two kind of players as if they are the same kind of players.

"Introducing" new Gleba mechanics is all dumped almost back-to-back, giving new players little time to adjust and learn the mechanics.

Speedrunner has to deal with the Gleba factory potentially failing and costing them the attempt, as time is a limited currency. See my next post.

When did I mix the two categories of players up? I am asking because I clarified several times that they aren't the same, and I want to improve my communication.

You're switching between them often enough that I have a difficult time parsing specifics between the two paradigms. I get the impression that you're making a complaint about one under the guise of the other. I don't think you're doing so intentionally, just that it's how it appears to me.

What inconsistency? You made it sound like I played Space Age for 2k hours when I didn't make that claim.

I responded to skepticism about how much of an 'expert' I am with Factorio by pointing out that I have thousands of hours of gameplay with Factorio.

I did not mention how much time I spent with Space Age, partially because there isn't a reliable way to track time spent playing with Space Age. Steam treats Factorio 1.1, Factorio 2.0 without Space Age, and Factorio 2.0 with Space Age as a singular number of total hours spent.

Since its release date, I have had at most a hundred hours of free time to play Space Age at my best guess, and even then, I didn't spend all of that time on Space Age.

Those statements are not mutually exclusive, and thus, there is consistency.

You complain that it simply takes too much time to master an expansion that effectively or factually more than doubles the content of Factorio that's only been public since late October. You spent a few thousand hours mastering vanilla and you're complaining because you can't master a part of this massive expansion on your specific terms in whatever weekend time you can cobble together in what amounts to three months now. That's an easy inconsistency to identify.

It doesn't matter why or how the blueprint/factory failed. The point that I am still trying to get you to understand is that when it does, it fails. It is frustrating that I spent hours doing whatever it was at the time only to find something failed or not working and not knowing why. See my next post.

Then I started to follow the trail of how something failed (no agricultural science, no flux, the particularities don't matter here), and it turns out that something went wrong with the Gleba Factory. Production charts only show what was being produced but not the particularises in how it failed or ceased producing.

The damage was already done by then, and I tried to revert to an earlier save game. However, I still couldn't find the cause for the failure as the earlier Gleba save game continued well past the fail timeframe without incident.

That is not the first or last time I have had to diagnose what went wrong with Gleba without data or even the only situation I have encountered, but it illustrates that dealing with Gleba is a bottomless well of dissatisfaction.

I tried empathizing. I offered to look at your blueprints. Hell, I'd be willing to look at your save. But you seem hellbent on complaining and not learning. Good luck with your speedrun, though I doubt you'll ever get it at this rate.

Let's review why I made that flippancy comment and what I was trying to do.

No. Let's not.

You decided it was okay to stoop to personal insults. I don't particularly care why. I gave you an offer that I'd look past such conduct, but perhaps I too heavily implied rather than specified that said offer required the minor concession of the point regarding personal time and it didn't even require public acknowledgement from you but simply keeping quiet about it.

You won't succeed justifying your comments in this regard because you can't. I recommend a more constructive use of your time involves focusing more on those many competing priorities you mentioned earlier, because you show no interest in a productive discussion. You spend an awful amount of time complaining about wasting time to people offering help.

[Mskvaer]

Waow 95% players can't leave Nauvis, and people complain about Gleba x)

Possibly some of them have not gotten the SA DLC and can not leave the planet.
I strongly suspect that the stats are from day one, when achievments were initially enabled, so again; lots of people that possibly stopped play Factorio before SA came out.
(But I'm just guessing)

[XT-248]

Look, speedrunning is an exercise in complete mastery of a game. If you are worried that your gleba factory breaks you need to get better at gleba, not have the game made easier for you. Spoilage is entirely deterministic, theres no randomness fucking you over. If your approach to not do bacteria yields suboptimal results find a better one. Routing and practice is a major part of speedrunning. My Gleba base has been running for over 30 hours and aside from eggs even restarted smoothely from a blackout.

As far as I know and in this thread, no one is asking to make Gleba 'easier.'

There is a well-known saying: "Easy to learn, hard to master."


I will expand on that saying in a way people can understand.

Gleba is "Hard to learn, hard to troubleshoot, and hard to master."



My problem is not that I am trying to get better at setting up a factory on Gleba. It is the lack of feedback from "what should have happened vs. what went wrong" and how to improve on the factory's flaws to prevent them from happening again.

Not understanding what went wrong initially makes the troubleshooting experience with Gleba very frustrating and less fulfilling.


Also, replacing suboptimal bacteria production with a reliable and optimized source of iron/copper from Nauvis, Vulcanis, or Space Platform results in a far more optimized process. I can stop/pause on demand without having to worry about complete gridlock from spoil or some other scenario that can result in a completely stalled Gleba factory (see my next post for an example of a Gleba factory failing without a clear cause).


The last thing I want to comment on is that the trick with a successful speedrun is that I don't need to be the world's top speedrunner (only those who want to be on the leaderboard do), and all I need to do is learn the right tricks that vastly improve my completion time. My first serious attempt on a speedrun ran past 8 hours, and then I started looking up tricks to improve my time-to-completion.

That is how I earned my original "There is no spoon run" on my second try by not "mastering" Factorio more than what I had already mastered in a few hundred hours of gameplay time in Factorio and using the right tricks. The requirement to "master" everything mechanic-wise for a successful Factorio speedrun is over-exaggeration.

These are good tricks to know: Press Z while dragging the mouse to move coal quickly between multiple furnaces, and hold Ctrl while clicking on furnaces with iron/copper plates inside. Additionally, aiming for early construction bots within the first two hours can significantly accelerate construction and expansion throughout the speedrun.

[XT-248]

Post Script:

While writing those posts, I was running a new design for a factory on Gleba in the background while not watching it.

I was using the following recipes: Agriculutural Science Packs, bioflux, yumako processing, jellynut processing, pentapod eggs, and nutrients from bioflux; all through bioreactors.

Guess what?

All I found was spoilage backing up and no damage to any part of the factory (turrets to prevent freshly hatched pentapods for bioreactor/science-pack-production from attacking and destroying stuff).

*snipped for brevity*

When I have had a bad day at work and want to come home to take my shoes off and relax, troubleshooting a Gleba factory is not my idea of a fun and relaxing afternoon.

Again.... you should consider asking for help. Please post in Gameplay Help and we'll help you get Gleba figured out. There's no shame in not figuring the whole thing out all by yourself. Posting in here over and over again how you can't figure it out and it's the most terrible game ever made because you can't get it, and nobody else can figure it out either, isn't going to help you get to Aquilo. Obviously, many thousands of people have already figured it out, and several of us are reading these forums, ready to help.

Respectfully, you think I need help where I may or may not need it.


I will post a screenshot of the failed Gleba Factory production chart and an overall look at the factory layout I mentioned last week.

I already stopped using this particular Gleba factory and started from scratch with a new redesign. So, any layout feedback or otherwise on the old Gleba design will not apply to the current Gleba design.


I have a couple of pieces of information to retain as you review this screenshot.

Rocket fuel from jelly/bioflux bioreactors, in the blue highlight area, is not running and has it own separate jelly bioreactor (second jelly bioreactor to the right).

The red highlight area completely crashed and somehow ran out of nutrients when it was not supposed to. The source of nutrients is bioflux-> nutrient, and can be seen in the middle of the red area and above the Pentapod Egg Bioreactors.

The production ratio I was going for with productivity module II Q1 in all of the bioreactors but for agricultural science, and if there is a decimal, then it is rounded up in the design: 3.5 agricultural science pack (extra bioreactor up to 5 and without productivity modules for crafting speed reason), 3.1 bioflux bioreactors, 2.3 yumako processing, 0.9 jelly processing, 9.7 pentapod eggs (one reactor is out of view), and precisely one bioflux->nutrient bioreactor.

The bioflux/nutrient for rocket fuel hasn't been added yet and will not leech from the above except for yumako/jellynut from farming.

I planned to have at least one agricultural tower for each Gleba agricultural product: yumako and jellynut. The Yumako agricultural tower can keep up with the above production, only needing half of the maximum capacity and having enough left over for rocket fuel production. Jellynut processing only needs an agricultural tower to run at 20% of maximum capacity.

I marked the production chart with two yellow arrows which is when I started running the Gleba factory at 100% capacity and walked away for a few hours. Then I saved the map/game, and exited Factorio.

At the first red arrow in the production chart, I started up Factorio and loaded the savegame from the second yellow arrow. Then I left it running in the background for a couple of hours until I saw the Gleba completely stalled out of the corner of my eye. I rushed over and saw the production line at the third red arrow on the bottom.

The only clue I have as to what could have gone wrong is an abnormally higher-than-normal spike in agricultural science production (see second red arrow), but that can't be because it was stable for at least an hour before that point.

02-01-2025, 16-59-04.png (1.62 MiB) Viewed 358 times

The pentapod egg bioreactors are almost but not quite 10 in the production ratio (9.7 rounded to 10 and average out to a ~3% underproduce of pentapod eggs), and agricultural science production spiked by at least 43% using eggs that don't exist somehow while I wasn't looking.

When I highlight the spike with my mouse in the production chart, it shows that the Gleba factory was somehow producing 191 agricultural science packs a minute (191/133 = 1.43 higher than expected, hence 43%). The average production rate was 133 per minute flat for at least a few hours. Only some of agricultural science bioreactors are supposed to be running at a time to get 133 agricultural science packs per minute, but somehow, the 4th and 5th came online at full speed when they were not supposed to. No logic or circuit is throttling agricultural science production as it is supposed to be self-regulating. The extra two agricultural science bioreactors help smooth the imbalance and fluctuation between bioflux and pentapod eggs.


One more thing to remember is that I haven't even started experimenting with stopping and starting the production line with this specific design (for when Nauvis labs are not using Agricultural Science packs for an extended period as previously mentioned).

The rocket silo completely removes all of the agricultural science packs as soon they arrive and sends them out to a space platform. The space platform deletes the science packs as soon they arrive (overthrowing into deep space or whatever). The rocket silo has access to unlimited ingredients for rocket parts (editor chest cheat). Power is taken care of off-screen and is overkill for this factory.


That is all the information you and I have to troubleshoot this at the moment. Good luck, as you will need it to figure out what went wrong.

If you think you can figure it out with some extra information. I might consider providing it if you can make a good case.

[XT-248]

When I initially tested some defense strategies on Gleba, I found that very little to nothing about defending against biters/worms was applicable to Gleba natives.

A defense perimeter of turrets and walls? Stomper walks over the wall and potentially becomes aggressive due to the presence of turrets.

Strafer spawns wriggler on top of the turrets and, as the namesake implies, strafe instead of moving in close.

I didn't ask for new enemies to act the same way as Nauvis fauna. I accepted and dealt with Vulcanis Demolishers of various sizes, for instance.

Yet, at the same time, Gleba's fauna is different enough that new players will have problems adjusting after Nauvis.

No, you want them to behave same as Nauvis from that checklist. You don't have complaints about Demolishers because you don't defend your factory from them. Vulcanus is all-offense. Nauvis teaches you what the basic tools are, how to keep them running, managing enemy spawners in the factory area to stay ahead of evolution, placement of defenses for interlocking fire... Do you even landmine? It's okay to be frustrated that you're struggling with a solution to a particular Factorio problem. You used the blueprints of others to facilitate a speedrun, so I'm not sure why you don't just simply look up Gleba defense strategies.

I didn't say that. You said it.

There is a disconnection between acknowledging that Gleba Fauna doesn't act the same way as Nauvis Fauna, and wanting Gleba and Nauvis Fauna to be the same/identical.

I said one of them but not the other.

The one that wasn't mentioned is inferred and does not help move the conversation in a constructive direction.

You're switching between them often enough that I have a difficult time parsing specifics between the two paradigms. I get the impression that you're making a complaint about one under the guise of the other. I don't think you're doing so intentionally, just that it's how it appears to me.

I am making the point that it is difficult for new players to learn Gleba's mechanics after dealing with production lines that do not resemble Gleba's production lines.

I am also making a separate point that when the Gleba production lines cease producing, finding out the reason/cause is problematic from a troubleshooting point of view when almost everything turns to spoilage.


Those two separate, distinct points do not overlap in any way or form. They can coexist simultaneously, and both are true.

See the above post for an example of a Gleba factory failing without a clear cause and illustrate that I am not a "novice," but I realize that people new to Factory Simulation Game might have a hard time with Gleba.

What inconsistency? You made it sound like I played Space Age for 2k hours when I didn't make that claim.

Those statements are not mutually exclusive, and thus, there is consistency.

You complain that it simply takes too much time to master an expansion that effectively or factually more than doubles the content of Factorio that's only been public since late October. You spent a few thousand hours mastering vanilla and you're complaining because you can't master a part of this massive expansion on your specific terms in whatever weekend time you can cobble together in what amounts to three months now. That's an easy inconsistency to identify.

No, the difference is that you think you identified an inconsistency where one may or may not exist and then treat the "inconsistency" as factual.


I spent thousands of hours of gameplay time over the years: I bought Factorio in 2016. To be more specific, it is 2k thousand hours of gameplay spread across nine years, which averages out to ~4.2 hours a week*.

*Math: 2,000 hours of gameplay divided by ( nine years * 52 weeks in a year ) = 2000 / ( 9 * 52 ) = 2000 / 468 = ~4.2 hours per week.


Which scenario seems more plausible: I spend thousands of hours in a few months, or over nine years, I spend playing a few hours each week?

I tried empathizing. I offered to look at your blueprints. Hell, I'd be willing to look at your save. But you seem hellbent on complaining and not learning. Good luck with your speedrun, though I doubt you'll ever get it at this rate.

When did you empathize with me? You have a funny way of showing it.

I am providing constructive feedback on Gleba's mechanics, and having others treat it as something other than constructive feedback doesn't help move the conversation forward meaningfully.


I have been experimenting with different setups for Gleba and learning without the help of the Factorio community.

I improved my Gleba design last week and then again just today, but they are both incomplete.

There is still a non-zero chance that the current Gleba redesign factory might fail like the previous design because I don't know how the previous one failed. See the above post.


It doesn't matter if you believe me or not.

Let's review why I made that flippancy comment and what I was trying to do.

No. Let's not.

You decided it was okay to stoop to personal insults. I don't particularly care why. I gave you an offer that I'd look past such conduct, but perhaps I too heavily implied rather than specified that said offer required the minor concession of the point regarding personal time and it didn't even require public acknowledgement from you but simply keeping quiet about it.

You won't succeed justifying your comments in this regard because you can't. I recommend a more constructive use of your time involves focusing more on those many competing priorities you mentioned earlier, because you show no interest in a productive discussion. You spend an awful amount of time complaining about wasting time to people offering help.

I am doing the review for your benefit.

I originally made a post as a reply to your post that was purely neutral in tone and respectful.

Your response afterward was then and continues to be uncalled for, and I am sure you know how to respond to posts more reasonably civilized than what happened.




Your enjoyment of Gleba's mechanics indicates that I should also enjoy them, even though fun is ultimately a subjective experience.

The truth is straightforward: not everyone finds the same things enjoyable.

[mmmPI]

Not understanding what went wrong initially makes the troubleshooting experience with Gleba very frustrating and less fulfilling.

That sound like you need help, it was appropriate to suggest gameplay help.

Respectfully, you think I need help where I may or may not need it.

You definitely seem to need it, given that you posted a setup and asked for help to troubleshot it just after :

That is all the information you and I have to troubleshoot this at the moment. Good luck, as you will need it to figure out what went wrong.
If you think you can figure it out with some extra information. I might consider providing it if you can make a good case.

That would be easier to provide help in the dedicated section, and if you provide the savefile.

No, the difference is that you think you identified an inconsistency where one may or may not exist and then treat the "inconsistency" as factual.

I have to agree @Jay_Raynor there, it's an inconsistency to claim what is problematic for new players "but not for you because you have played already thousand of hours" but you still wants something changed "for the new players". Because it's hard to learn, but at the same time when you're not a new player you don't need to learn anymore, and when you're a new player you pretty much expect to have to learn new things.

The truth is straightforward: not everyone finds the same things enjoyable.

That's why there are so many mods !

I recommended one to make the game easier last time, because it still seem to me that it is what you describe with your word, you need more time to troubleshoot your factory, to understand what went wrong because you're not familiar with the new mechanics, there are mods to slow down the game !

[Muche]

So, any layout feedback or otherwise on the old Gleba design will not apply to the current Gleba design.
[...]
That is all the information you and I have to troubleshoot this at the moment.

I would disagree. The information you have and we don't is the actual layout and underlaying assumptions leading to it.
Some of them could be gleaned from the design implementation. You shared some of the assumptions in the abstract.
How did they materialize after the incident? Did spoilage get onto some belt where it was not supposed to be? Were all belts/machines empty (because all spoilage was properly disposed of)? Are storage levels in buffer chests (if present) as expected?

Yes, investigating the incident after half an hour is harder, as short lifespan items have spoiled.
What about adding some alarms to warn when some of these assumptions break, to allow for immediate investigation?

You shared the production graph.
I can see some bumps at about 0.5 h mark in consumption of some items.
There is a slight bump in science production at 2 h mark. As this is after over 1 h of stable production, can it be somehow correlated to 1h spoilage time of fruits?
Are they expected / related?
The difference between total production and consuption of mash, jelly, and Yumako suggests some of them are spoiling. Are all parts of factory able to deal with it, or are some parts where it's assumed nothing will ever spoil here?

So when you say layout feedback will not apply to later designs, it sounds like you're treating those assumptions as correct and not willing to wave goodbye to some of them.
Maybe some of them are affected by bugs, like 125825 Inserter Deadlock when nutrients spoil in Inserter hand?

[angramania]

That is all the information you and I have to troubleshoot this at the moment. Good luck, as you will need it to figure out what went wrong.
If you think you can figure it out with some extra information. I might consider providing it if you can make a good case.

Looks like you intentionally hide necessary information. Screenshot is nothing, you should provide blueprint. And even screenshot is limited, we do not see a lot of entries from production/consumption report and what items were on belts. Still having this little information I would suppose that you have forgot about seeds. Their consumption is much less than production. As result at some moment there was no place for inserters to remove seeds, processing of yumako and jellynut have stopped and everything have collapsed in a few minutes.

[Tertius]

My problem is not that I am trying to get better at setting up a factory on Gleba. It is the lack of feedback from "what should have happened vs. what went wrong" and how to improve on the factory's flaws to prevent them from happening again.

A good approach for Gleba is to not let something bad happen in the first place. Don't build factories under the assumption it will always work with the common case, and if something fails you will identify the issue and fix that manually. No, it's extremely tedious "just build", then "find the error". Instead, design a proper production line in the first place with all edge cases in mind, even the unlikely ones, even with the "impossible" in mind. Design it for the edge case: everything spoilable spoils in an instant. From one tick to the other. Every production line must be able to recover from that.

That means:
- no belt that carry spoilables must end somewhere. Instead, it has to loop somehow.
- if you use direct insertion, one inserter per machine must also be present to output spoilage to the spoilage recycling system
- belts must never stall. The loop must keep looping at all times. If a belt with spoilables stall, and in one location is an inserter to insert a spoilable item to a machine, and the items on the stalled belt spoil, this inserter doesn't have an item to input, so this machine stalls. If this is an important machine, the whole factory can stall.


You mentioned the ore bacteria. It's very straightforward to design a failsafe bacteria production line. You start with one biochamber with the bacteria recipe. The next biochambers in line get the cultivation recipe, all inserting from and outputting to the same belt as the first machine. The end of that belt has a mixture of spoilage (from the first machine), bacteria (from all machines) and ore (if something stalls and bacteria spoil). So you filter spoilage with a filtering splitter and add the rest to chests to let the bacteria spoil to ore. To stop the first chamber producing stuff while the real line is working, connect it with the second chamber and read content including in crafting. Activate the first only if bacteria=0.
That's all.

There is this video to visualize this approach (load in separate window to have controls to start/stop/rewind):

out.mp4

(21.14 MiB) Downloaded 9 times

It starts from everything clogged or being empty. I just connect input and output, and everything initializes itself. It's also able to restart if ore isn't being used for a longer amount of time (I accelerated this and just showed the restart).

It's also showing my approach to Gleba: design in the map editor instead of directly on Gleba. You have additional tools, and you can do the most debugging in a safe environment. This way, the blueprints I actually build on a real map need only final finetuning but it's already proven the design works in general. And if it didn't work out, I go back into the map editor and develop required changes there.

[XT-248]

I would disagree. The information you have and we don't is the actual layout and underlaying assumptions leading to it.
Some of them could be gleaned from the design implementation. You shared some of the assumptions in the abstract.

I shared the factory and the production chart in the abstract because I abandoned this design without knowing why it failed.

Which is the whole point of this exercise.

How did they materialize after the incident?

What do you mean exactly by 'materialized'?

When I was away from the keyboard, bioflux, pentapod eggs, and nutrients bioreactors were constantly running throughout the test without any external interference.

How would I know if any of them materialized or were produced naturally?

Did spoilage get onto some belt where it was not supposed to be?

Not that I could find. But then again, if there is a chance of it spoiling, there is an outlet. There are inserters filtered only to remove spoil on both bioreactors and belts everywhere. I seriously doubt that happened.

Were all belts/machines empty (because all spoilage was properly disposed of)?

Yes, all the inserters and bioreactors were empty (input, outlet, and fuel alike in bioreactors), with a few bioreactor still with intact 1-hour fruits and intact pentapod eggs. The only spoil items I could find were on the left-most belt that goes off to the left past the production chart near the split jelly bioreactors. That goes to a bioreactor that self-start with some nutrients and a constant stream of spoil and churn out nutrients. Those nutrients go straight to yumako and jellynut processing bioreactors and nowhere else.

Are storage levels in buffer chests (if present) as expected?

The only logistic chests that interacted with the red factory area are the blue/requester logistic chests that requested Yumako and Jellynut. They only requested a small number, just enough to keep them running. As you can see in the screenshot, there are no flying logistic drones. The seeds, from processing, were placed in the logistic chests, and none were full or backed up.

The few non-fruit requests were all low levels of nutrients (10 at the most) from spoilage to ensure that yumako and jellynut processing bioreactors are supplied with nutrients no matter what.

Yes, investigating the incident after half an hour is harder, as short lifespan items have spoiled.
What about adding some alarms to warn when some of these assumptions break, to allow for immediate investigation?

Where would you suggest placing those alerts? What should I check/monitor for?

You shared the production graph.
I can see some bumps at about 0.5 h mark in consumption of some items.

An abnormal amount of spoil was consumed 11 minutes before the abnormal spike/raise in agricultural science pack production (what you called the 0.5-hour mark).

02-02-2025, 01-59-18.png (733.7 KiB) Viewed 189 times

My Gleba factory is supposed to get rid of spoilage. Yet there was virtually zero spoil consumed between the 2.5-hour marker and the 0.5-hour marker, and bam, 350 spoil per minute showed up while I was away and sporadically flip-flop between low consume and high consume after that point until the entire Gleba factory ceased.

The thing is, I am unable to explain how the Gleba factory continued for a few minutes without any incident until agricultural science packs died out after ~12 minutes. The length of time rules out jelly and yumako mash (4 and 3 minute timers, respectively). It can't be nutrient either, as that has a 5-minute timer. It could possibly be pentapod eggs sitting idle on the belt, except there are no replacements that occurred during this timeframe and would, in all likelihood, damage the inserter/belt that allowed a large amount of spoil to leave the Gleba factory by design.

There is a slight bump in science production at 2 h mark. As this is after over 1 h of stable production, can it be somehow correlated to 1h spoilage time of fruits?
Are they expected / related?
The difference between total production and consuption of mash, jelly, and Yumako suggests some of them are spoiling. Are all parts of factory able to deal with it, or are some parts where it's assumed nothing will ever spoil here?

I have verified that fruits and their byproducts stopped being consumed at least 10 minutes, or 0.2 hours after agricultural science production failed. I fully expect that some of it will be spoiled since I cannot obtain the whole number of bioreactors needed in the provided production ratio earlier, with a single exception that I have previously mentioned (bioflux to nutrient bioreactor). Even then, the Gleba factory over-produces nutrients with the spoil recipe, whose excess goes straight to fruit processing.

02-02-2025, 02-12-04.png (1.04 MiB) Viewed 189 times

So when you say layout feedback will not apply to later designs, it sounds like you're treating those assumptions as correct and not willing to wave goodbye to some of them.
Maybe some of them are affected by bugs, like 125825 Inserter Deadlock when nutrients spoil in Inserter hand?

That is a real possibility.

Except I only saw one occurrence of this hours before the cease of the Gleba factory production lines. It was on the last pentapod egg bioreactor's outlet inserter (on the very bottom in the first screenshot with the factory in the abstract) and wouldn't have interference with the bioflux-to-nutrient bioreactor in the center in a very destructive manner.




P.S.

I just realized the nutrient/bioflux production/consumption 1-hour chart might be useful here. Remember, yumako mash and jelly production were fine for at least 10 minutes after the 0.5-hour time marker. So the raw ingredients for bioflux didn't run out. So, in theory, bioflux production should have been fine.

02-02-2025, 02-29-39.png (655.5 KiB) Viewed 189 times

So, I expanded my search for data to all Gleba fruit-related items with the agricultural science pack failure at the 15-minute marker.

02-02-2025, 02-43-00.png (986.1 KiB) Viewed 189 times

The same chart, as above, but extended to 10-hour time frame.

02-02-2025, 02-58-05.png (874.99 KiB) Viewed 189 times

Those production charts raise even more questions than answers, especially seeing that there were multiple abnormalities that occurred during the 2.5 and 0.5 time markers that didn't result in a cataclysm downspiraling.

Spoil spike occurred at least three minutes before any of the other abnormal spikes in bioflux/jelly (bioflux is the orange top line with two twin mountain-like peaks, and jelly is the smaller red line that peaked only three time above itself before trending downward at the same time as the others).

Why was there a huge sudden burst of spoil consumption right then at the 0.5-hour marker? Where did it come from? How did it get in the system all at once? Before any of the other abnormalities in bioflux/jelly production?

The consumption rate was steady throughout the hours I spent away from the Gleba factory, raising even more questions than answers.


So, I am back to square one without a new lead to investigate what happened to Gleba Factory that I haven't considered.



Also, my current work-in-progress Gleba factory is like this now and far more stabilized at a higher production rate, but it requires more material to set up than I wanted, which is why I am looking to downsize. Some option includes but not limited to removing beacons and expanding the bioflux production line. I still haven't made the necessary modifications for it to self-restart from scratch.

02-02-2025, 09-53-13.png (3.25 MiB) Viewed 189 times

[XT-248]

My problem is not that I am trying to get better at setting up a factory on Gleba. It is the lack of feedback from "what should have happened vs. what went wrong" and how to improve on the factory's flaws to prevent them from happening again.

A good approach for Gleba is to not let something bad happen in the first place. Don't build factories under the assumption it will always work with the common case, and if something fails you will identify the issue and fix that manually. No, it's extremely tedious "just build", then "find the error". Instead, design a proper production line in the first place with all edge cases in mind, even the unlikely ones, even with the "impossible" in mind. Design it for the edge case: everything spoilable spoils in an instant. From one tick to the other. Every production line must be able to recover from that.


That means:
- no belt that carry spoilables must end somewhere. Instead, it has to loop somehow.
- if you use direct insertion, one inserter per machine must also be present to output spoilage to the spoilage recycling system
- belts must never stall. The loop must keep looping at all times. If a belt with spoilables stall, and in one location is an inserter to insert a spoilable item to a machine, and the items on the stalled belt spoil, this inserter doesn't have an item to input, so this machine stalls. If this is an important machine, the whole factory can stall.

See my previous post about a sudden burst of spoil out of nowhere.


Why is it quite difficult for people to understand that I am anything but a person new to Factorio? And I really don't need help, but thank you anyway.

You mentioned the ore bacteria. It's very straightforward to design a failsafe bacteria production line. You start with one biochamber with the bacteria recipe. The next biochambers in line get the cultivation recipe, all inserting from and outputting to the same belt as the first machine. The end of that belt has a mixture of spoilage (from the first machine), bacteria (from all machines) and ore (if something stalls and bacteria spoil). So you filter spoilage with a filtering splitter and add the rest to chests to let the bacteria spoil to ore. To stop the first chamber producing stuff while the real line is working, connect it with the second chamber and read content including in crafting. Activate the first only if bacteria=0.
That's all.

There is this video to visualize this approach (load in separate window to have controls to start/stop/rewind):

out.mp4
It starts from everything clogged or being empty. I just connect input and output, and everything initializes itself. It's also able to restart if ore isn't being used for a longer amount of time (I accelerated this and just showed the restart).

It's also showing my approach to Gleba: design in the map editor instead of directly on Gleba. You have additional tools, and you can do the most debugging in a safe environment. This way, the blueprints I actually build on a real map need only final finetuning but it's already proven the design works in general. And if it didn't work out, I go back into the map editor and develop required changes there.

See my example of the Agricultural Science Pack for 10 hours in my previous post.


At an initial glance, I would need 1k red belt, almost 400 modules, 500 inserters, enough laser turrets to cover the Agricultural Science Production line, enough iron plates and copper plates to craft things locally, etc.

Each item requires a rocket launch, which costs valuable and limited time. I am open to suggestions on how to substantively downsize this factory. Not adding to it.


The Agricultural Science row is four times longer copy/paste, which isn't visible in the screenshot.

Also, notice how all of the yumako mash and jelly are either directly inserted or not on the external main bus. To have a bacteria production would require another refactor and even more materials to build the bacteria production lines + jelly/yumako-mash inputs.

Notice how my bioflux-to-nutrient bioreactor (top row of machines next to pentapod eggs and agricultural science pack bioreactors) moves nutrients to the buffer/green chest? Some of that goes to the bioflux bioreactor production to the blue chest without input display on the right side (between bioflux outlet and the topmost spoil outlet). That is one of the large refactors I made, which would have invalidated most of the feedback based on the previous design, as it wasn't doing that before. If I hover over the logistic network, I can see that it has 1k nutrients in stock ready to be picked up, which should help mitigate the odds of running out of nutrients in my previous design.

I forgot to label the rocket fuel outlet, but it is there between bioflux input and water input. So that is all of the energy and rocket silo upkeep covered.

[Muche]

How did they materialize after the incident?

What do you mean exactly by 'materialized'?

When I was away from the keyboard, bioflux, pentapod eggs, and nutrients bioreactors were constantly running throughout the test without any external interference.

How would I know if any of them materialized or were produced naturally?

I meant how the assumptions have materialized. That is, if a belt is supposed to be empty after items spoiling, is it indeed empty after the incident, or is there some spoilage on it?
You describe them later though.

Yes, investigating the incident after half an hour is harder, as short lifespan items have spoiled.
What about adding some alarms to warn when some of these assumptions break, to allow for immediate investigation?

Where would you suggest placing those alerts? What should I check/monitor for?

Did spoilage get onto some belt where it was not supposed to be?

I seriously doubt that happened.

Add an alarm if it gets there.

Are storage levels in buffer chests (if present) as expected?

The only logistic chests that interacted with the red factory area are the blue/requester logistic chests that requested Yumako and Jellynut. They only requested a small number, just enough to keep them running. As you can see in the screenshot, there are no flying logistic drones. The seeds, from processing, were placed in the logistic chests, and none were full or backed up.
The few non-fruit requests were all low levels of nutrients (10 at the most) from spoilage to ensure that yumako and jellynut processing bioreactors are supplied with nutrients no matter what.

Alarms for the levels of fruits in the chests.
The amounts in the logistic network.
Whether the machines are always running, if they are expected to.
Seeds levels (when empty and when full).
Nutrients level for these critical chests/machines.

There was virtually zero spoil consumed between the 2.5-hour marker and the 0.5-hour marker, and bam, 350 spoil per minute showed up.

An alarm when spoilage belt is too empty or too full. Or calculate item throughput and check that.

Generally, an alarm at any place about which you say "should not happen/must be present at all times".


Fruit production graphs show repeated fluctuations (Yumako at 2.5, 1.5, 0.5, Jellynut at 3, 2, 1).
As there are no similar flutuations in their consumptions, it suggests the spoiling occurs between trees and fruit consumption machines.
How does that part of factory cope with bursts of fruits spoiling?

Spoilage blockage lasting for an hour would suggest some errant fruit blocking spoilage belt. Two hours points more to an errant science pack or bioflux.
However, your new design seems to utilize bots, not belts.
If that's what you meant by abandoning (the belt) design, then yes, feedback/improvements about belt handling might not be directly applicable to bot handling.

Another possible source of information is kill stats and/or entities produced/build, in case of pentapods attacks/hatching.
I would expect any losses to be replaced almost immediately.
Which is another assumption that needs to be checked. Either manually from graphs, or automatically via alarms.

[Tertius]

See my previous post about a sudden burst of spoil out of nowhere.

There's some detail you might miss about nutrients, spoilage and biochambers. Spoiled nutrients can accumulate invisibly during normal operation in a separate spoilage slot in addition to the regular temporary trash output slots. They can accumulate till there is one full stack (200) of them. Only then, if the 201th spoilage is created by a spoiled nutrient, the machine blocks until the spoilage is output.
If a biochamber gets not so fresh nutrients or stalls intermediately for a short amount of time, so nutrients spoil, they accumulate in a separate trash slot. It can happen this slot isn't being emptied immediately but instead the machine continues working as soon as new nutrients get inserted. For example, in one production line I set the output inserters to use spoilage priority and grab "freshest first". This left spoilage resulting from nutrients in place as long as there was regular output, and since the machine was continuously producing output fresher than spoilage, spoilage was never extracted until the slot was full.
Might perhaps be the cause of your failure.

I wanted to grab "freshest first" to get seeds from yumako and jellynut mashing. Without that, I only got the mash and never the seeds until a full stack of seeds accumulated. Since the seeds don't spoil, "freshest first" got me seeds first, then mash, then spoilage. But since there was always new mash in the output slots, any spoilage was never removed.

Why is it quite difficult for people to understand that I am anything but a person new to Factorio? And I really don't need help, but thank you anyway.

We are all new to Gleba's spoilage mechanics and its intricate details. We all need to explore new methods to deal with that. All your (and my) thousands of hours played are irrelevant. Get off your high horse.

See my example of the Agricultural Science Pack for 10 hours in my previous post.

With a few more strategically placed speed beacons, you could achieve vastly more output with vastly less footprint. You're transporting too much spoiling items over too large distances. You even store nutrients - what a waste. It spoils. Instead, create nutrients on the spot, feed the nutrient biochambers with their own nutrients, and kickstart them with a spoilage-to-nutrient assembling machine which turns off as soon as there are nutrients > 0 in the nutrient biochambers.

I'm not yet ready with my complete Gleba factory, but the science-related modules are this.
The first one produces bioflux, a certain amount of nutrients for various independent modules not shown, and other yumako/jellynut byproducts. It kickstarts from whole fruits (the farms run with electricity only) and spoilage-to-nutrients. It's a 2-level kickstart. First it kickstarts yumako mashing, then yumako mash to nutrients, and this finally enables jellynut mashing and the first bioflux chamber, so the real bioflux-to-nutrient production starts.

02-02-2025, 18-44-02.png (1009.48 KiB) Viewed 145 times

And this is the science production.
It's supposed to be directly plugged behind the bioflux production. It has its own nutrient production, since it uses direct insertion due to the high nutrient throughput not feasible for belts. It just needs a few nutrients from the first step to start its own nutrient production.
A small gimmick is present in the bottom left, it's a keep-alive for a single pentapod egg in case something goes wrong with bioflux supply or stuff stalling. As soon as the nutrient machines get empty, a single egg is put into a chest and refreshed every 10 minutes with nutrients from spoilage and put back once the regular nutrient machines start again. During keepalive, any pentapod eggs are expunged into the spoilage exit belt to keep pentapods appearing from egg spoiling. Only the eggs within the chambers will stay and spoil, but I limit their number. I thought about recipe changing to expel them as well, but didn't find the time yet. And pentapod egg production doesn't really need production modules, the effect is minor.

02-02-2025, 18-49-49.png (1.17 MiB) Viewed 145 times

If it comes to performance, this is what it created during my development. The spikes are from tests when I removed supply, electricity or deliberately let the factory stall.
I guess 639/min with about 95% fresh science packs is more than the huge production line you showed. 2nd image has ~50 modules and 170 belts, 1st image has 100 modules and ~220 belts. Both together exactly 20 beacons, which is one rocket stack.

02-02-2025, 19-03-13.png (140.93 KiB) Viewed 145 times

ps. as far as I see, the Gleba general bootstrap is this:
- import an assortment of intermediates
- create 2 biochambers
- create a yumako farm with 1 agricultural tower
- import heating towers
- create seeds and artificial soil to make a proper farm, as well as a huge amount of spoilage (2-3 chests full)
- create a jellynut farm with 1 agricultural tower
- create seeds and artificial soil to make a proper farm
- create a preliminary pentapod egg factory
- create a bunch of biochambers with imported ingredients
- start a bioflux production line
- start 2 bacteria production line to get iron and copper ore
- create iron/copper processing with a few foundries (with that ship from Vulcanus, also import a bunch of green belts)
- create a mall to be independent of imports
- create the rest of the factory from stuff produced locally
- start science production

[NineNine]

ps. as far as I see, the Gleba general bootstrap is this:
- import an assortment of intermediates
- create 2 biochambers
- create a yumako farm with 1 agricultural tower
- import heating towers
- create seeds and artificial soil to make a proper farm, as well as a huge amount of spoilage (2-3 chests full)
- create a jellynut farm with 1 agricultural tower
- create seeds and artificial soil to make a proper farm
- create a preliminary pentapod egg factory
- create a bunch of biochambers with imported ingredients
- start a bioflux production line
- start 2 bacteria production line to get iron and copper ore
- create iron/copper processing with a few foundries (with that ship from Vulcanus, also import a bunch of green belts)
- create a mall to be independent of imports
- create the rest of the factory from stuff produced locally
- start science production

Or, you can just land on the planet and start with nothing. That's what I did.