Sokomine reports: lag-death. all was frozen. aft ... #4124
Labels
No Label
1. kind/balancing
1. kind/breaking
1. kind/bug
1. kind/construction
1. kind/documentation
1. kind/enhancement
1. kind/griefing
1. kind/invalid
1. kind/meme
1. kind/node limit
1. kind/other
1. kind/protocol
2. prio/controversial
2. prio/critical
2. prio/elevated
2. prio/good first issue
2. prio/interesting
2. prio/low
3. source/art
3. source/client
3. source/engine
3. source/ingame
3. source/integration
3. source/lag
3. source/license
3. source/mod upstream
3. source/unknown
3. source/website
4. step/approved
4. step/at work
4. step/blocked
4. step/discussion
4. step/help wanted
4. step/needs confirmation
4. step/partially fixed
4. step/question
4. step/ready to deploy
4. step/ready to QA test
4. step/want approval
5. result/cannot reproduce
5. result/duplicate
5. result/fixed
5. result/maybe
5. result/wontfix
ugh/petz
ugh/QA main
ugh/QA NOK
ugh/QA OK
No Milestone
No project
No Assignees
5 Participants
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: your-land/bugtracker#4124
Loading…
Reference in New Issue
No description provided.
Delete Branch "%!s(<nil>)"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Sokomine reports a bug:
Player position:
Player look:
Player information:
Player meta:
Log identifier
Profiler save:
Status:
Teleport command:
Compass command:
The recorded max_lag does not hint at the server having issues.
Might be clientside or connection issue
there certainly were a few huge lag spikes, cf. #4123
disabling the anticheat mechanism might reduce the incidence of lag deaths by not putting a player back into a deadly situation that they thought they ran away from.
i wish there was a way to turn the anticheat on/off on the fly - e.g. during events and when lag exceeds some threshold
Could have been anticheat. But that'd have been only one contribution then.
How it looked to me: Mobs and players were frozen (animation still running).
Before, I had tried to kill a Small. It wasn't the least impressed. No change to his HP, though (with very huge delay) some hitting sounds.
Earlier incarnations of lag like this usually ended after a couple of seconds. Not this time. It remained frozen. The server (or client) just wasn't particulary intrested in what I was doing. I could still run around, but feedback from the server was not arriving. Eating golden apples worked, but with terrible delay. Eventually, while moving alone with nobody around, i got hit with a few seconds delay and couldn't do anything to it but watch. Then I saw my bones appear. After that I could still walk around for 5-10 seconds until I got the death screen.
And after respawning, my character was turned randomly around for at least 30 seconds. Extremly annoying. Had to look elsewhere.
Chat worked to some degree, although with considerable delay compared to IRC.
My connection is usually very stable and reliable. Admittedly, it is very slow by today's standards.
Some network packages regarding movement definitely got out of sync or lost. There's the random turning at the end, so my movements 1-2 minutes earlier probably got processed then (after respawning).
Summary of what Alias and I found out through testing: Each mob moving around sends a data packet. If there are too many of them, my internet connection is overflooded. MT apparently doesn't have adequate handling for such a huge amount of mobs.
the fact that the server step is so short might mean that more packets are sent every second, if mob positions are updated on every step.
Perhaps we could disable smooth rotation (rotation in 8 steps) for voice mobs? It only affects random walking mode. But then - not all are always attacking, and any packet not sent is welcome.
As long as mobs don't do anything, there's most likely no new packet sent.
But even an update every second by about a 1000 mobs would certainly be too much.
Imagine you have 1000 mobs who want to send you a packet each step. One step is 0.47ms, that's about 21 steps per second. That's 21 000 packets per second. With a 1 Mbit/s landline you get at max 125 Kilobyte per second. 125 000 Bytes / 21 000 packets means, a packet may have less than 6 Bytes of overall content. Even the header is larger.
... and Sokomine bytes the dust.
sry couldnt resist
possibly relevant upstream issue: https://github.com/minetest/minetest/issues/10694
this seems to be controlled by a "global" setting for mobs_redo, and not per-mob: https://notabug.org/TenPlus1/mobs_redo/src/master/api.txt#L772
i vote we disable it and see if anything improves?