I disagree.
While failed hops happen all the time because a lot of networks block ICMP, they can still show an issue when the trace fails to reach its destination.
And you can most definitely have your ISP resolve failed hop issues if its within the first few hops / within their network of control. (I've done this multiple times)
Sometimes it is just packet fragmentation, network congestion, a link is out, the server took a dump, etc.
That being said, I think you are right in this case, it looks like it means nothing. I hadn't noticed before that it successfully reached it's destination.
@myx those pingtests look good
I'm definitely stumped.
I would start investigating software installed and settings. If you have another computer that you could try on the same network, that would be a great test
Yes, the same result on laptop. I connected to Wooza server and get black screen.
log.txt: https://yadi.sk/i/WHtaTTvncCzMd
The same result on my PC. But if I turn on VPN, all ok, no black screen, no lags.
This is strange, but with standart MTU 1492, maximum length of ns2 UDP packet was 1442. Then I changed MTU to 1380, and still get 1442 length UDP packets. WTF?
Or maybe i made mistake? Packet size in Wireshark call "length"?..
That's strange indeed but I suppose it didn't solve the problem. I'm observing similar behavior here, maybe I misunderstand something and MTU doesn't work like that with UDP.
You should probably install OpenVPN to your PC or use a custom firmware like dd-wrt or OpenWRT for the router. Then you need to setup a server with OpenVPN somewhere or use a free/paid VPN provider. There are plenty of them.
I don't use VPN in router. I use it directly from PC. Just open OpenVPN site, there you can find nice vpn service, that called "Private Tunnel". 100Mb traffic for free. Enough to test VPN. I know, sounds like advertisement, but it not ^^
IronHorseDeveloper, QA Manager, Technical Support & contributorJoin Date: 2010-05-08Member: 71669Members, Super Administrators, Forum Admins, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Blue, Subnautica Playtester, Subnautica PT Lead, Pistachionauts
The bottom line is that no matter what the server shouldn't be sending empty packets..
Regardless of your network configuration.
Like I said one of our engine programmers made that error in lieu of crashing. I'll talk to him and see if there's a way to capture more data to diagnose this
Could all list their ISP name, Router model number, modem type, Operating system, download and upload speeds, whether you're wired or wireless, and whether this happens on every server or just specific ones - it will help further diagnose this issue.
Thanks!
Sometimes on Russia # and YOClan.co.uk servers, I don't play much on other servers, played on The Thirsty Onos a few times but all was OK.
It has a very strange behavior, someday I can play YOClan.co.uk server perfectly without any lag/plug, other day the YOClan.co.uk server can lag for me all the day, so I have to change server. @IronHorse
After the 267 build the Server browser is working weirdly for me. It shows Favorite & last servers instantly, but not all other servers. The server counter shows 0. Also I can't update the server's info. The update button works time to time, sometimes I have to restart the game to make it work. Here is video where server browser works yet reasonably:
And no, it's definitely not the ISP problem. It began when I was on ADSL ISP, now I have other ISP with fiber connection.
P.S.
Yes, I know the Wooza's servers can't be reliable on diagnosing the network problems, but people massively play on these servers, and mb it can be interesting info.
As always, Red plugs and MainThread : Error: OnPacket: Received an empty NetworkPacket.
I had a severe problems, other players from Russia also had problems but still playable.
Here is the link to log with net_log 3
IronHorseDeveloper, QA Manager, Technical Support & contributorJoin Date: 2010-05-08Member: 71669Members, Super Administrators, Forum Admins, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Blue, Subnautica Playtester, Subnautica PT Lead, Pistachionauts
@METROID
Yeah its no surprise that you'd have issues in that server with that playercount and that abysmal tick rate - when the server is on average performing less than half of what it's supposed to be you're going to run into issues.
It's a poor place to test this EmpyPacket problem. It can create a false positive.
As for people massively playing on them : They are given a warning when they join so they know what is occurring.
After some review of this thread, the only video that's helpful is the YO clan server video
Could you put together more videos like that please? I am in talks with the person who runs it, so hopefully we can get some server logs as well and compare.
Also both @Myx and @METROID Could i have both your Techsupport zip files please?
Thanks guys!
IronHorseDeveloper, QA Manager, Technical Support & contributorJoin Date: 2010-05-08Member: 71669Members, Super Administrators, Forum Admins, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Blue, Subnautica Playtester, Subnautica PT Lead, Pistachionauts
I know it's happening on typical servers @ATF I even said it was - I need more videos of those.
I am sure larger servers produce this issue easier (makes sense given the networking choke and demands) but there's just too many other factors with them - like terrible tick rates - to be a reliable test case.
@METROID regular tech support, not when the game is open.
OK, I'm will continue to post my struggles. Now it's 270 build. Overall I began to warp sometimes, or see other players warping.
1. Server: ^cV.fi - Competitive Gaming
I connected to server and played ns2_jambi flawlessly. All you can see in the provided log file. Then map changed to Biodome, at first it was OK, but on the minute 6 the nightmare came ))) The first thing was the warping skulk at 5-6 minute, and on minute 9 everything was warping, so the situation was gradually getting worst.
LOG file (net_stats 3): https://dl.dropboxusercontent.com/u/36046461/NS2/log_cv_server.zip
2. Server: Russia #1
I can say that Russia #1 - always being a number 1 non-problematic server on the Russian-speaking community.
But now sometimes I see warping players, I'm warping too. Also, I have a strange feeling that skulk's bites sometimes do not register properly.
LOG file (net_stats 3): https://dl.dropboxusercontent.com/u/36046461/NS2/log_russia1_server.zip
Some timestamps:
0:11 - Server RED!
1:09 - Server YELLOW!
4:46 - Server YELLOW!
6:10 - Server RED!
6:12 - Server YELLOW!
7:02 - Marine warped away.
7:30 - Server YELLOW!
8:36 - Server RED!
9:40 - Server RED! & I warped.
10:22 - Server YELLOW!
12:28 - Server YELLOW!
13:52 - Server YELLOW!
I have new question to developers: how working ingame ping? Because I have strange result on one server. Ingame ping rushes between 60 and 300ms. But if i open cmd.exe and ping it as usual, I get normal result. I hope this information will be useful for my ISP.
IronHorseDeveloper, QA Manager, Technical Support & contributorJoin Date: 2010-05-08Member: 71669Members, Super Administrators, Forum Admins, Forum Moderators, NS2 Developer, NS2 Playtester, Squad Five Blue, Subnautica Playtester, Subnautica PT Lead, Pistachionauts
edited December 2014
This is regarding the latency disparity only:
(so as not to be confused with the red plug issue in general)
Assuming you don't have similar antivirus or firewall software and settings on your laptop (confirm this please? because it really matters. A fresh install of windows would be best) I still have to assume that it is your ISP at fault here, if a separate computer (your laptop) and bypassing the router all yield the same results.
This means it's got nothing to do with software on any of the PCs or Router - this includes NS2, because its software.
However it could be that particular socket type (that is incredibly common!) that your ISP has issues with. Have you tried other games that use UDP? Teamfortress 2 uses the same exact UDP socket type and default port: 27015
Please try pinging TF2 servers in steam
No any antivirus and firewall software. Windows reinstalled few months ago at a time when I bought SSD.
Now the morning and I have no problems with ping to this server. Tonight I try it again, when the workload is high. By the way, I want to emphasize that such a strange ping only on this server, others have stable ingame ping always.
I will install TF2 and test it tonight. But I spent some time playing "Insurgency" that based on Source engine. I didn't noticed any lags or strange ping on few servers that i played.
I have already written in support of my ISP with last video included. ATF told me, that ns2 use UDP packets for ping requests. So, is it correct, if I add message about poor performance of UDP into ISP's support request? Any advice on how to test the UDP and put the result with the evidence for my ISP?
Thanks for your support.
By the way, I want to emphasize that such a strange ping only on this server, others have stable ingame ping always.
Any advice on how to test the UDP and put the result with the evidence for my ISP?
If you have access to the server you can use PsPing in latency + UDP mode to make measurements.
Hi again. I wanna bump this thread, still I have this problem. And just wanna share log.txt.
I was a spectator on the server, sometimes I saw a red plugs. After few minutes I got long red plug and probably game wanted to to kick me because of the timeout (i saw "Main Menu Initialized at Version: 275" in log.txt) But it didn't happened, instead this i got strange messages in console. And disconnected by "disconnect" command.
Not sure if this is related to my problem, but I think that game behavior is not supposed to do. Sorry for broken english
[ 52.531] MainThread : Error: OnPacket: Received an empty NetworkPacket
^same thing we're getting before redplugs,
Best thing i'd advise right now is to disable any mods you have enabled (from what it looks like you might have a weapon modification? (other then crosshair etc) and re-check your ns2 integrity.
No any antivirus and firewall software. Windows reinstalled few months ago at a time when I bought SSD.
Now the morning and I have no problems with ping to this server. Tonight I try it again, when the workload is high. By the way, I want to emphasize that such a strange ping only on this server, others have stable ingame ping always.
I will install TF2 and test it tonight. But I spent some time playing "Insurgency" that based on Source engine. I didn't noticed any lags or strange ping on few servers that i played.
I have already written in support of my ISP with last video included. ATF told me, that ns2 use UDP packets for ping requests. So, is it correct, if I add message about poor performance of UDP into ISP's support request? Any advice on how to test the UDP and put the result with the evidence for my ISP?
Thanks for your support.
For the record, I don't have those empty packets anymore. At least, according to the latest log. There are the infamous "Attempting to sync an invalid collision rep" errors, they have timestamps so I see I played 7600 seconds, more than two hours. No red plugs, no empty packets. I only remember red plugs appearing when the server crashes but it's expected, of course.
Scanning UDP ports won't reveal anything, UDP is a stateless protocol, there's no "connection" possible. ISP may mark your UDP traffic as low-priority and even drop it in case of internal congestion. That's perfectly fine in protocol terms as UDP isn't reliable but handling massive drops is a tedious task and there may be issues in the game itself.
BTW, I see that Myx connects via PPTP which is a VPN of sorts. This introduces a whole lot of possible issues. For example, a friend of mine has a very interesting issue with a single game, Star Ruler 2. When he joins my game his internet connection just drops instantly. Probably because of UDP flood when the game state is transferred to him. The same happens when he hosts the game and I connect to him, he just disconnects from the internet completely and has to wait until the router reconnects. We solved this issue from my side, I've configured fire-qos traffic shaper (frontend for Linux tc) to limit the outgoing traffic to his IP to 2 Mbps, then everything was fine. It's important to note that he also uses PPTP while I'm on PPPoE. He couldn't connect directly without the router because of some issues between his Ubuntu and that Russian PPTP type. I didn't even know that local PPTP types exist but they are — on some routers it's mentioned explicitly, like, compatible with Russian PPTP.
So I think the best possible solution at this point is to switch to a PPPoE ISP. After all, PPTP was developed by Microsoft, remember when they did something good? Me neither, hehehe. Seriously though, it's a terrible protocol that incapsulates traffic into GRE and ALSO uses a side TCP channel for control, I don't even want to know what was inside the heads of those who invented this. OpenVPN and L2TP (if I'm not wrong) are perfectly fine with UDP and a single port for both control and payload traffic.
Comments
While failed hops happen all the time because a lot of networks block ICMP, they can still show an issue when the trace fails to reach its destination.
And you can most definitely have your ISP resolve failed hop issues if its within the first few hops / within their network of control. (I've done this multiple times)
Sometimes it is just packet fragmentation, network congestion, a link is out, the server took a dump, etc.
That being said, I think you are right in this case, it looks like it means nothing. I hadn't noticed before that it successfully reached it's destination.
@myx those pingtests look good
I'm definitely stumped.
I would start investigating software installed and settings. If you have another computer that you could try on the same network, that would be a great test
log.txt: https://yadi.sk/i/WHtaTTvncCzMd
The same result on my PC. But if I turn on VPN, all ok, no black screen, no lags.
http://steamcommunity.com/app/4920/discussions/0/613938693091594355/
http://steamcommunity.com/app/4920/discussions/0/613935404092491754/
So i think this problem not mine only. Second topic remind me, that after 267 build i had same problems with crashes and red plug. If i changed team, or take commander place, i got red plug for 5-10 seconds and then crash. Now i have red plug under the same conditions, but no game crash. Maybe roots of this problems somewhere near?
Regardless of your network configuration.
Like I said one of our engine programmers made that error in lieu of crashing. I'll talk to him and see if there's a way to capture more data to diagnose this
Connection type: fiber (protocol PPTP)
Router: ZyXEL Keenetic Giga II
OS: Windows 8.1 x64 (all updates)
Speed: 60/60mbps
Wired
Sometimes on Russia # and YOClan.co.uk servers, I don't play much on other servers, played on The Thirsty Onos a few times but all was OK.
It has a very strange behavior, someday I can play YOClan.co.uk server perfectly without any lag/plug, other day the YOClan.co.uk server can lag for me all the day, so I have to change server. @IronHorse
P.S.
As always, Red plugs and MainThread : Error: OnPacket: Received an empty NetworkPacket.
I had a severe problems, other players from Russia also had problems but still playable.
Here is the link to log with net_log 3
Yeah its no surprise that you'd have issues in that server with that playercount and that abysmal tick rate - when the server is on average performing less than half of what it's supposed to be you're going to run into issues.
It's a poor place to test this EmpyPacket problem. It can create a false positive.
As for people massively playing on them : They are given a warning when they join so they know what is occurring.
After some review of this thread, the only video that's helpful is the YO clan server video
Could you put together more videos like that please? I am in talks with the person who runs it, so hopefully we can get some server logs as well and compare.
Also both @Myx and @METROID Could i have both your Techsupport zip files please?
Thanks guys!
OP and others have reported those errors on multiple servers.
Regardless of how the 42 player server performs, try to avoid using it to test for this issue. It could create false positives.
So, you need an ordinary Techsupport.zip, or when the problem occurs?
I know ))) Just for additional info.
I am sure larger servers produce this issue easier (makes sense given the networking choke and demands) but there's just too many other factors with them - like terrible tick rates - to be a reliable test case.
@METROID regular tech support, not when the game is open.
Tech_support:
YO.Clan servers works fine for me, so I'm not sure that I will able to make good video...
1. Server: ^cV.fi - Competitive Gaming
I connected to server and played ns2_jambi flawlessly. All you can see in the provided log file. Then map changed to Biodome, at first it was OK, but on the minute 6 the nightmare came ))) The first thing was the warping skulk at 5-6 minute, and on minute 9 everything was warping, so the situation was gradually getting worst.
LOG file (net_stats 3): https://dl.dropboxusercontent.com/u/36046461/NS2/log_cv_server.zip
2. Server: Russia #1
I can say that Russia #1 - always being a number 1 non-problematic server on the Russian-speaking community.
But now sometimes I see warping players, I'm warping too. Also, I have a strange feeling that skulk's bites sometimes do not register properly.
LOG file (net_stats 3): https://dl.dropboxusercontent.com/u/36046461/NS2/log_russia1_server.zip
Some timestamps:
(so as not to be confused with the red plug issue in general)
Assuming you don't have similar antivirus or firewall software and settings on your laptop (confirm this please? because it really matters. A fresh install of windows would be best) I still have to assume that it is your ISP at fault here, if a separate computer (your laptop) and bypassing the router all yield the same results.
This means it's got nothing to do with software on any of the PCs or Router - this includes NS2, because its software.
However it could be that particular socket type (that is incredibly common!) that your ISP has issues with.
Have you tried other games that use UDP? Teamfortress 2 uses the same exact UDP socket type and default port: 27015
Please try pinging TF2 servers in steam
Now the morning and I have no problems with ping to this server. Tonight I try it again, when the workload is high. By the way, I want to emphasize that such a strange ping only on this server, others have stable ingame ping always.
I will install TF2 and test it tonight. But I spent some time playing "Insurgency" that based on Source engine. I didn't noticed any lags or strange ping on few servers that i played.
I have already written in support of my ISP with last video included. ATF told me, that ns2 use UDP packets for ping requests. So, is it correct, if I add message about poor performance of UDP into ISP's support request? Any advice on how to test the UDP and put the result with the evidence for my ISP?
Thanks for your support.
If you have access to the server you can use PsPing in latency + UDP mode to make measurements.
I was a spectator on the server, sometimes I saw a red plugs. After few minutes I got long red plug and probably game wanted to to kick me because of the timeout (i saw "Main Menu Initialized at Version: 275" in log.txt) But it didn't happened, instead this i got strange messages in console. And disconnected by "disconnect" command.
Not sure if this is related to my problem, but I think that game behavior is not supposed to do.
Sorry for broken english
https://yadi.sk/d/ik5mtpuIitZLC
Be careful, log.txt 30Mb.
^same thing we're getting before redplugs,
Best thing i'd advise right now is to disable any mods you have enabled (from what it looks like you might have a weapon modification? (other then crosshair etc) and re-check your ns2 integrity.
http://www.speedguide.net/scan.php
Try this
Scanning UDP ports won't reveal anything, UDP is a stateless protocol, there's no "connection" possible. ISP may mark your UDP traffic as low-priority and even drop it in case of internal congestion. That's perfectly fine in protocol terms as UDP isn't reliable but handling massive drops is a tedious task and there may be issues in the game itself.
BTW, I see that Myx connects via PPTP which is a VPN of sorts. This introduces a whole lot of possible issues. For example, a friend of mine has a very interesting issue with a single game, Star Ruler 2. When he joins my game his internet connection just drops instantly. Probably because of UDP flood when the game state is transferred to him. The same happens when he hosts the game and I connect to him, he just disconnects from the internet completely and has to wait until the router reconnects. We solved this issue from my side, I've configured fire-qos traffic shaper (frontend for Linux tc) to limit the outgoing traffic to his IP to 2 Mbps, then everything was fine. It's important to note that he also uses PPTP while I'm on PPPoE. He couldn't connect directly without the router because of some issues between his Ubuntu and that Russian PPTP type. I didn't even know that local PPTP types exist but they are — on some routers it's mentioned explicitly, like, compatible with Russian PPTP.
So I think the best possible solution at this point is to switch to a PPPoE ISP. After all, PPTP was developed by Microsoft, remember when they did something good? Me neither, hehehe. Seriously though, it's a terrible protocol that incapsulates traffic into GRE and ALSO uses a side TCP channel for control, I don't even want to know what was inside the heads of those who invented this. OpenVPN and L2TP (if I'm not wrong) are perfectly fine with UDP and a single port for both control and payload traffic.
This should be fixed for the next patch. https://trello.com/c/HrkdyDcI/785-network-protocol-issues-bugs-in-lag-loss-handling
Hope some of you report back on its success afterwards