Hazeron Forums
2022-03-10 Server Load - Printable Version

+- Hazeron Forums (https://hazeron.com/mybb)
+-- Forum: Shores of Hazeron (https://hazeron.com/mybb/forumdisplay.php?fid=1)
+--- Forum: Updates (https://hazeron.com/mybb/forumdisplay.php?fid=11)
+--- Thread: 2022-03-10 Server Load (/showthread.php?tid=2671)

Pages: 1 2 3 4 5


RE: 10-03-2022 Server Load - Celarious - 03-15-2022

(03-14-2022, 09:07 PM)Deantwo Wrote: Your other ideas are probably good and all. But they assume that the game just stays the way it is, when clearly the root cause of the problem is that players are abusing the current systems to cause way more server load than expected. It is always a good idea to try and change player behavior to cause less server load, and if that change also gives players incentives to cooperate more it is just an extra plus.

I doubt it. Unless you genuinely believe we somehow had a part in hundreds of thousands of systems erroneously loading (and never unloading) on day 3, there is clearly a severe server-side bug causing far more systems to be active than should be. All of your suggestions work under the assumption that people did, indeed, send out 50,000 survey ships to scan a hundred thousand unique systems, which quite literally isn't possible given the avatar and officer counts, even today, let alone during the first week. Regardless, even if they did, systems shouldn't be kept loaded if no colony is established and are not regularly visited. It's also not happened anywhere nearly as bad in far more popular previous universes, which rules out players entirely.

Let's not blame the players for a pretty obvious issue in the back-end.

If you want to generally talk about players doing less server-slowing things like dropping a million electronic units on top of each other, sure. But there's no reason to think that's the cause of the issues described by Haxus and others in this thread and thus is not relevant to this discussion, should be in a separate thread, so let's focus on the big glaring issue bringing the servers to their knees


RE: 10-03-2022 Server Load - Deantwo - 03-15-2022

(03-15-2022, 02:38 AM)Celarious Wrote: Let's not blame the players for a pretty obvious issue in the back-end.

Haxus didn't mention that they were generated in error, so I assume they weren't. I will say that the number of generated solar systems is massive, but we already know that common player behavior is the maximize the amount of solar systems they visit per hour. And low-tech travel further increase this.

There are empires that sent out ships with orders to just visit whole sectors while they are AFK/sleeping, just so they can then have a survey ship lazily fly between already explored solar systems and scan them one at a time.

Deadheading, unlike warp, travels through all solar systems in their path. So if empires sent out ships towards the galactic core, assuming dense sectors that could easily be 10 solar systems generated every hour per empire.

I guess it could also be bugs that cause solar systems to generate when they aren't needed. Maybe new avatar spawning generates solar systems at random until it fines a habitable one? Maybe the new intergalactic wormholes at stellar black holes generate random solar systems in the destination galaxy to find an exit wormhole? Or maybe pirate base spawning is generating solar systems at random near players to find valid locations?

We don't know what type of bug could possibly be causing it, so trying to guess is kinda pointless. I am sure Haxus will find the cause if they are generated by a bug on his own. Therefore discussing the methods that we do know about makes more sense, hence suggestions that could lower the amount of solar systems generated during surveying for resources.

As for why solar systems aren't "unloaded" a few hours after they are empty? How would anyone know why Haxus hasn't done that already? I only have a wiki edit Haxus made explaining when solar systems decay. So maybe the thought of "unloading a solar system" never occurred to him or maybe it wasn't a viable solution in the long run, none of us here will know unless Haxus decides to tell us. However arguing that a suggestion is stupid/unnecessary because of a solution that might not be viable/possible could potentially exist seems counter productive.


RE: 10-03-2022 Server Load - QuakeIV - 03-15-2022

Would all need to be averaging one new system explored every 4.8 seconds or so (read: EVERY, both day and night) since the start of the universe, for some back of the envelope type math. Seems faintly absurd to me.


RE: 10-03-2022 Server Load - Celarious - 03-15-2022

(03-15-2022, 07:26 AM)Deantwo Wrote: Haxus didn't mention that they were generated in error, so I assume they weren't. I will say that the number of generated solar systems is massive, but we already know that common player behavior is the maximize the amount of solar systems they visit per hour. And low-tech travel further increase this.

There are empires that sent out ships with orders to just visit whole sectors while they are AFK/sleeping, just so they can then have a survey ship lazily fly between already explored solar systems and scan them one at a time.

Deadheading, unlike warp, travels through all solar systems in their path. So if empires sent out ships towards the galactic core, assuming dense sectors that could easily be 10 solar systems generated every hour per empire.

I think you're not really grasping just how many systems 150 THOUSAND is. On the very third day, when a lot of empires (by their own admission) didn't even have basic spacecraft, there were this many active. But let's assume all 501 avatars existed on day 3 and all of them had been deadheading 24/7 through unexplored sectors since the universe started. As Quake said, every avatar would need to have been visiting a system every few seconds, which just isn't possible with lightspeed, even with galactic core sector density. Nor would it be possible had we all have been using warp 1 since the start either, since as far as I know that abstracts a bit and doesn't load every system in the path. Regardless, you've ignored the fact that people would've been doing the exact same thing in previous universes, in fact even moreso as you could argue the TL system gave an incentive to colonize more worlds, and yet those weren't even remotely as bad after a week.

(03-15-2022, 07:26 AM)Deantwo Wrote: We don't know what type of bug could possibly be causing it, so trying to guess is kinda pointless. I am sure Haxus will find the cause if they are generated by a bug on his own. Therefore discussing the methods that we do know about makes more sense, hence suggestions that could lower the amount of solar systems generated during surveying for resources.

You're also guessing when you suggest that players are the problem and that fixing the system generation from this angle has any relevance to the issue at hand. This also seems pointless by your logic, no? This angle also contradicts previous universes, so it doesn't make much sense either.

(03-15-2022, 07:26 AM)Deantwo Wrote: As for why solar systems aren't "unloaded" a few hours after they are empty? How would anyone know why Haxus hasn't done that already? I only have a wiki edit Haxus made explaining when solar systems decay. So maybe the thought of "unloading a solar system" never occurred to him or maybe it wasn't a viable solution in the long run, none of us here will know unless Haxus decides to tell us. However arguing that a suggestion is stupid/unnecessary because of a solution that might not be viable/possible could potentially exist seems counter productive.

We don't know why or even if they unload. What we do know is that previous universes with the same bugs and players didn't have this issue, and using everything Haxus and other players said in this thread, alongside numbers given in the standings, we're able to make an educated guess that players are most likely not the cause of this one. We have no evidence showing as such, and even if the cause *was* somebody finding a way to maliciously generate thousands of solar systems instantly, we currently have no reason to believe as such. That's why your suggestions aren't really super relevant here, they're more guesses than anything else. Haxus wouldn't have mentioned there being 270k systems "fast" if it was unrelated, and given that players are unable to have explored this quickly, server-side bugs are the obvious culprit. Blindly blaming the playerbase for this seems counter-productive.

PS: I didn't say your suggestion was unnecessary, just that it doesn't seem relevant to these specific issues and that we should focus on the more likely glaring causes. Regardless, if unloading unused assets is "not viable", then it needs to be made viable immediately since almost no games keep everything permanently loaded, as it's a huge waste of hardware resources as we can see here


RE: 10-03-2022 Server Load - AlrianneG - 03-15-2022

Quote:Servers have been very heavily loaded managing 270k solar systems.


270k systems activly simulated? If this is correct, it's absurd.


You need to simulate only where players and ships are !!


RE: 10-03-2022 Server Load - resonkinetic - 03-16-2022

(03-15-2022, 05:53 PM)AlrianneG Wrote:
Quote:Servers have been very heavily loaded managing 270k solar systems.


270k systems activly simulated? If this is correct, it's absurd.


You need to simulate only where players and ships are !!

I second this. That's such an unnecessary load.


RE: 10-03-2022 Server Load - drog007 - 03-16-2022

Good grief... when the worst of the lag hit and we heard about the sheer number of systems generated nobody on the leader board had more than 5 systems claimed. I find it incredibly unlikely that people had ALL been visiting SEVERAL ORDERS OF MAGNITUDE more systems than that. Obviously some mechanic triggered that generated them all, and whether it was expected or unintentional is entirely speculative at this point; but it very obviously wasn't direct player visits or scans.

In other words, it's entirely inappropriate to be pointing fingers at Haxus for "obvious issues," and similarly ridiculous to try be calling to curb player behaviour. Give Haxus time to ensure things are back up properly infrastructure wise, and an opportunity to look for issues. There's an entire forum section for game suggestions/ideas in the mean time.


RE: 10-03-2022 Server Load - AlrianneG - 03-16-2022

Quote:I find it incredibly unlikely that people had ALL been visiting SEVERAL ORDERS OF MAGNITUDE more systems than that.

I saw somewhere that the server generates all the systems in a sector and all the neighboring sectors of where a player is. With warp, the number number of generated system can rize really quickly.


RE: 10-03-2022 Server Load - Greydog - 03-16-2022

I think it Athena. She had each of her alts fire up one of those 100pc radars of hers ;)


RE: 10-03-2022 Server Load - Celarious - 03-16-2022

(03-16-2022, 09:51 AM)drog007 Wrote: Obviously some mechanic triggered that generated them all, and whether it was expected or unintentional is entirely speculative at this point; but it very obviously wasn't direct player visits or scans.

In other words, it's entirely inappropriate to be pointing fingers at Haxus for "obvious issues," and similarly ridiculous to try be calling to curb player behaviour. Give Haxus time to ensure things are back up properly infrastructure wise, and an opportunity to look for issues. There's an entire forum section for game suggestions/ideas in the mean time.

Exactly this. No fingers pointed at Haxus, he's a single person trying to manage the whole thing, hope he finds whatever it may be causing the excessive generation