São Paulo Gateway LAG

up

TEXAS Gateway is the way to go for me :\

Don't know if I should be happy or sad that I'm not the only one having problems. I didn't play 3.0 but just got back playing and there's some serious problems with fps drops and lag spikes. I used to run the game smoothly with not problems at all, and even though I know there were graphical improvements I don't see how my pc would struggle the way it is.

I mostly get dced while loading, even though they don't last not even 25-30 seconds. In every area I need to stay stil for 10 seconds for it to load and leave the 5 fps mark, I lowered the resolution and all the possible settings, the only thing keeping me alive is Dynamic Resolution.

Regarding the whole lag and disconnect issue, here is my WinMTR: https://pastebin.com/svt0UYdQ

I hope whatever is wrong with the game gets fixed, it's been so long since I started playing and I hate to experience the 10 acts for the first time with so many problems.

Vivo ISP, São Paulo Realm.

The same issue, every 20~ minutes got a HUGE spike and DC. It is funny that is affects Discord also - only my voice to the people, I keep hearing them normally. Whatsapp, Netflix, Twitch and etc feels the same, so it is not a problem with the ISP.

Today I tried playing on Texas server: Played for 2 hours, 0 spikes, 0 unexpected DC, 0 Discord robot voice.

It is something between Sao Paulo realm and some ISP routes, most notably VIVO.

Worth mentioning that after the last patch the game was smooth, and was using the same ISP, the same setup. Started in the lasted update for me.

Same issue here. Basically a huge spike then DC.


My WinTMTR report after 2-3 hours of gameplay catching the moment when this issue happened.

https://pastebin.com/X0JGiw3f

Looks like ae6.dar02.sao01.networklayer.com is the source of the problem as posted by this user here

"

awterado escribió:

This is getting insane, lost 3 ethernal lab runs because of this. For what I've seen, the ae6.dar02.sao01.networklayer.com stops responding for a few seconds then commes back. The WinMTR report won't give detail unless you start it as soon as the game freezes.

Same problem here... 15 20 minutes latency spike and DC (only SAO PAULO Gateway)

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 192.168.15.1 - 0 | 1505 | 1505 | 0 | 0 | 12 | 0 |
| No response from host - 100 | 302 | 0 | 0 | 0 | 0 | 0 |
| 187-100-179-42.dsl.telesp.net.br - 0 | 1505 | 1505 | 0 | 2 | 19 | 1 |
| 187-100-197-218.dsl.telesp.net.br - 0 | 1505 | 1505 | 1 | 2 | 38 | 7 |
| 187-100-34-129.dsl.telesp.net.br - 0 | 1505 | 1505 | 1 | 2 | 35 | 3 |
| 84.16.10.162 - 0 | 1504 | 1504 | 2 | 5 | 36 | 4 |
| 84.16.14.10 - 0 | 1505 | 1505 | 5 | 6 | 20 | 6 |
| 84.16.8.65 - 1 | 1481 | 1475 | 4 | 5 | 61 | 5 |
| ae6.dar02.sao01.networklayer.com - 1 | 1497 | 1495 | 5 | 7 | 61 | 6 |
| po2.fcr01b.sao01.networklayer.com - 1 | 1493 | 1490 | 5 | 7 | 58 | 7 |
| 36.84.39a9.ip4.static.sl-reverse.com - 1 | 1493 | 1490 | 5 | 6 | 58 | 6 |
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider

I solved my problem with Dc`s installing a VPN as suggested by support.
My connection is usually stable, but from the past week i started having lag spikes from 50 to 1700 followed by disconects.
I`m using Sao Paulo vivo isp.

I installed PIA VPN and that solved the disconnect problem but it is a bit laggy.

Does anyone knows a better VPN service to play POE in Brazil?

Hello,

I've been having the same problems since patch 3.1.1c hit (not saying it is the problem, just to mark a point in time), it could be earlier than that. The reason I am associating it to 3.1.1c is because right after the servers came back up that day, I logged on to play and died on Abyss HC because of the latency issue for the first time this league, I noticed it happens quite often since then.

What I see as a player, cycloning my way through mobs, character stops answering to commands and stay idle, but mobs continue to hit me, I see my life going down as the leech buffer from Endless Hunger is over, and then I am dead. So it seems that the server is still communicating until the regular disconnect time is reached (is it still 6 seconds ?). Using the logoff macro does nothing, it gets me to the login screen as expected, but the server still sees me as logged on, until the disconnect time is reached, sometimes this is fatal.

I've been playing in the Texas gateway for the last two days, this does not happen, but playing cyclone with 160ms (and thus using predictive mode) is disgusting.


What I see as a mainframe system programmer that used to work for IBM for decades (what I do to earn my money), with some knowledge on network and problem determination. I had a traceroute run for a few hours with nothing running on my nerwork besides the trace, no routers or devices connected, no other applications running, local firewalls and antivirus programs disabled. It seems to me, the very last hop (#8) before reaching GGG infrastructure has some losses and then the remaining hops also have losses (#9, 10 and 11), those belong to IBM's Softlayer if I am not mistaken, which is ultimately GGG's responsability.


https://pastebin.com/ZUmPxWUa

I want to point out that it seems odd to me that the 8th hop is 84.16.8.65, that IP address according to Whois is located in Spain, I doubt it would go all the way to Spain and back to Brazil and still maintain 20ms the entire path. Other than that, from my traceroute and from the others that I saw on this thread, it all seems to be related to the networklayer servers.

Under regular circunstances, losing packets is not a big deal, as long as they are not in sequence, which is exactly the case here, they all miss one after the other for a few seconds in a row, it can be seen in the image below that I drew from my traceroute.

https://imgur.com/a/Ujc4r

There are 10/11 errors for the last hops, they are all in sequence (the red bar in the graph). Everything else looks perfectly normal for the rest of the trace.

Again, I am not pointing fingers at GGG or anything, just trying to show some data that I gathered and hope this helps if the problem is indeed within their infrastructure, as I stated above, I am not a network engineer, not qualified to conclude anything.

It also shows a timestamp when this occured, if it helps to isolate the problem, would someone dare to take a closer look.

Thanks and sorry for the long post.

*Edit*
Just eddited the image on IMGUR, I forgot to stop the traceroute while I was posting this and it happened again at 16:06 BRT, so I posted another image on IMGUR with a 60s view of the trace when the spike happens, you can see about 10s without response from the last hops, and then a few hiccups, all good after that.

*Edit 2*
Left the traceroute running for a few hours, noticed this pattern where the latency spike occurs roughly every 35 minutes, added a new image to the album to show it:
https://i.imgur.com/ekzMqn2.jpg

"

bimonti escribió:

*sip*

I suggest you to post this on reddit.
This forum is a ghost town and we need both players and staff to be knowledgeable about this issue.

"

i8037 escribió:

Same issue here. Basically a huge spike then DC.


My WinTMTR report after 2-3 hours of gameplay catching the moment when this issue happened.

https://pastebin.com/X0JGiw3f

Looks like ae6.dar02.sao01.networklayer.com is the source of the problem as posted by this user here

"

awterado escribió:

This is getting insane, lost 3 ethernal lab runs because of this. For what I've seen, the ae6.dar02.sao01.networklayer.com stops responding for a few seconds then commes back. The WinMTR report won't give detail unless you start it as soon as the game freezes.

No, it shows the hop imedietly before the network layer has the problem not hte network layer 84.16.8.65 - 2 | 1920 | 1897 | 5 | 11 | 404 | 11