Implemented Disable /tps and make an auto-meter like AE lag meter.

/tps doesn't cause TPS issues, tps issues would still be there because of it, and it doesn't make that much of an impact.

The AE Lag meter thing is already horrendous because it makes people think "AE by default is laggy" when it really isn't. It gives players the wrong impression

 

AE by itself isn't that bad, it's when people push it's autocrafting capacities as far as they can because they lack patience that lag happens. Also I'm not sure about other servers, but, on infinity some players seem to take pride on how much lag they cause the server.

 

i have a suggestion for the new 1.12 servers, so recently sky adventures was running very very VERY well at 20 TPS for a solid week then out of nowhere something happened and the tps locked down to what stoneblock is, a solid 5 ~ 9 TPS and from what I've been hearing through the grapevine the admins didn't exactly set the servers up correctly, I.E - Missing vital plugins, - The servers were rushed out, - packs weren't looked at thoroughly and now there is lag beyond repair due to certain items not being disabled / banned from the start, - the servers are being hammered by intensive resource hogging mods like roost and refined storage interactions and other various non sense

this is by far the worst MyM has done so far to date but they can fix it regardless

my suggestion is this, run 2 servers instead of the 1 on either end as when you have reserve slots up to 30 / 18 , it can be very very problematic as most servers reserve ram per player dynamically and that can cause problems when there are too many players for that 1 server to handle

my other suggestion is to have all of the mods and admins pitch in to re-examine the packs, obviously there is much more at play that is causing some extreme problems for us average players to even try and think of playing and at a steady TPS of 5 ~ 9 the pack is virtually unplayable and thats when you have to question the stability of it

 

Just a reminder that this is the conditions you agreed to when joining the server

 

i am not being rude, i am simply stating my thoughts, my reasoning behind those thoughts and what i've heard, personally i feel like the servers do need to be looked at again, cause quite litterally a solid 20 tps server went to a locked 9 or below over night so something or someone did something it or she/he wasnt supposed to do and i've tried telling admins before that the roosts mod is just fubar'd as its coding is horrendous for servers as it throws random negative values and that if you dont empty them fast enough (the provided roost collector cant even keep up with its own mod ) they cause the server to go oh sh* and i've seen it first hand

 

i would say more like look at implementing something to check how many there are in certain chunks cause they do cause problems

 

i personally cannot code in java tho i've wanted to learn how, but from what i've seen is that there are problems that popped up over night so is just my opinion and thoughts

 

This seems contrary to what staff has been saying. If staff is correct and it causes issues by profiling on demand, then I suggest to instead modify the tool to correct this problem. Here's what I'm thinking:

• Staff can access the original on-demand profiling /tps command
• Everyone else when they run the /tps command will give the last generated report instead of re-generating it. Server can refresh this report as often as it likes maybe 10min intervals or something. Alternative 1, the tool could be modified to have lower impact profiling that runs continuously and aggregates the results every 10mins or something for the pre-canned report. Alternative 2, the tool could generate the report on-demand for the first person to issue it, and give that report to all other instances of /tps until a certain internal has expired (ie. rate limit the profiling). Either way the timestamp of when the report was last generated should be included in the message.

I like this solution over the report broadcast because either we need to make the interval extremely long (30mins) and not be able to differentiate network/blocklag vs tps lag in time to fix issues/check on farms before they become a problem; or we'd need to make it short (10mins) and upset people due to the server spam.

I agree that people tend to fixate too much on the AE lag meter as if it was the only source of lag. However, I wouldn't say the tool was horrendous. Many have found the current tool useful and has helped reduce the impact of AE systems before they start causing tps issues. I believe we are better off with it as-is than not at all. That said, if it were to be improved, here's a couple of thoughts about how to do that (one of the following):

• Only report systems that are actually having an impact on TPS drop (as much as is reasonably feasible to do); or
• Only send the reports to the players who have claimed the area where the offending ae system is located (server wide like normal if not claimed); or
• Generate the reports like normal, but do not broadcast the reports and instead let players access via command the last generated report (like my tps suggestion above) and maybe with a disclaimer

Anyways, IMO these are pretty low priority improvements, but if a dev is bored would be nice to haves

 

if ae was not laggy, then why would one need a lagg'0'meter

 

a lot of stuff in modded is laggy, it just happens that AE is very widely used, and it takes some work to use it without causing to much lag

 

Literally no one said that AE was not laggy.

 

It's not as laggy as new people think it is because of the lag-o-meter, I never said it wasn't laggy if you had a massive setup with autocrafting, but that it isn't as bad as people think it is.

 

You also have to keep in mind, AE, even with massive auto crafting setups can be lag free (to a degree) if you sub network and use p2p tunnels. Everytime my AE builds get to a certain size with auto crafting I always sub network and/or put p2p tunnels on it. On DJ, my ME system showed up there once at a rating of 5 because I was auto crafting a 16384k cell for a quest. I had 20+ interfaces full of patterns and a dozen more scattered around my base for auto smelting, blood magic, even went out of my way and connected some to EnderIO fluid tanks to automate lava and water buckets for certain recipes that needed them.

 

just wanna drop this here

Performance Guide — MyM-Wiki 0.0.1 documentation

 

well heres the thing when i was on infinity with my very first 1.7.10 modded AE system i caused the server to crash with it automating all the things about 3 years ago and when i finally understood why, it took me a better half of 6 hours of re-working my ae system with only the drives connected with over 2,000 channels branched from it, AE by itself isnt laggy at all, the lagg o meter is just a visual representation of the systems on the server causing the most hits to the tps, i had one that litterally crashed it out due to lag, so when you get ALL THE THINGS going it can hurt the server pretty quickly