Jump to content


UPDATE: Minsk Devs Respond to Server Lag


  • Please log in to reply
364 replies to this topic

dance210 #241 Posted May 13 2017 - 16:58

    Community Specialist

  • -Players-
  • 24318 battles
  • 4,451
  • [D4NCE] D4NCE
  • Member since:
    05-31-2012

View PostMongoose_warrior, on May 12 2017 - 06:12, said:

I am pretty well fed up with lag issues. I have been playing 3.5 yrs and the last 3 months have been frustrating. I play a game at 61 fps and the next at 2 fps. It really kills my fun. I have checked everything, isp, PC, setup, etc. No change. This is very unsatisfactory.

 

View PostSquirty69, on May 12 2017 - 11:13, said:

Why won't they admit what it is? Hmmmmm. Mea culpa means they loose money in the form of refunds. Hmmmmm. 9.18 started the lag problem for me. Geforce 660 with driver 382.05, 8 gig ram, Amd fx 8300, windows 10 64 bit, Xfinity 60 megabit connection. Central IL. avg. 45 ms ping on East server, 75 on west. I run between 85-115 fps until the lag spike, it then drops to 30-45. No rhyme or reason at all. HD client, already un/reinstalled. Turned off all audio devices not in use under windows device manager. Turned "multi-display/mixed-gpu acceleration" to single display performance mode under the geforce driver. Nothing works. I don't trust WG on this. I've tried all advice that I could find so far.

 

Hi guys!

 

For anyone who is having lag / server / ping issues - please do send us a Support ticket. We are aware of the situation and are collecting specific information to help diagnose and hopefully address this issue.

 

For the Support ticket, we need:

  • PingPlotter run while issues are occurring
  • WG Check while issues are occurring
  • Replay showing the lag

 

You can go here to find a nice guide on what is needed to report the issues.

 

All three items should be attached to the Support ticket. If there are any questions, please let me know.



ImABadTealREEEE #242 Posted May 13 2017 - 18:26

    Corporal

  • -Players-
  • 9125 battles
  • 88
  • Member since:
    12-13-2014
is it worth it any compensation for our lost premium time or anything?

rich73 #243 Posted May 13 2017 - 18:42

    Major

  • Players
  • 56023 battles
  • 6,488
  • Member since:
    10-17-2011

View Postdance210, on May 13 2017 - 07:58, said:

 

 

Hi guys!

 

For anyone who is having lag / server / ping issues - please do send us a Support ticket. We are aware of the situation and are collecting specific information to help diagnose and hopefully address this issue.

 

For the Support ticket, we need:

  • PingPlotter run while issues are occurring
  • WG Check while issues are occurring
  • Replay showing the lag

 

You can go here to find a nice guide on what is needed to report the issues.

 

All three items should be attached to the Support ticket. If there are any questions, please let me know.

 

You know this is just a traffic related event and there is nothing anyone can do about traffic.With more and more people streaming nowadays,the infrastructure is stressed and only upgrades in the area where the bottlenecks are, will fix anything.I myself was rerouted for almost 3 years to west server,resulting in a 130 ping versus a 20 to 30 ping I enjoy now.I had to play east at this time,and I live in the west, as the ping was better,funny thing being that east never changed the whole period but did get better when I was grandfathered to a better internet plan.The better internet plan had no effect on west.So internet traffic is affected by routing and this changes as there is more traffic and switching comes into effect,resulting in being routed a farther distance,which gives worse performance.So rather blame streaming and insufficient upgrading of the infrastructure,tho if W G can improve their performance somehow to maybe compensate somehow,that would be welcome too.

badmonkey59 #244 Posted May 13 2017 - 18:56

    Captain

  • Players
  • 44017 battles
  • 1,414
  • [-_W_-] -_W_-
  • Member since:
    05-15-2011

View PostImABadTealREEEE, on May 13 2017 - 11:26, said:

is it worth it any compensation for our lost premium time or anything?

 

lol. This question will not be answered. I have been asking since March when the trouble started.



ImABadTealREEEE #245 Posted May 13 2017 - 19:13

    Corporal

  • -Players-
  • 9125 battles
  • 88
  • Member since:
    12-13-2014

View Postbadmonkey59, on May 13 2017 - 18:56, said:

 

lol. This question will not be answered. I have been asking since March when the trouble started.

 

FR dude i mean i want my premium time back

ImABadTealREEEE #246 Posted May 13 2017 - 19:15

    Corporal

  • -Players-
  • 9125 battles
  • 88
  • Member since:
    12-13-2014

View Postdance210, on May 13 2017 - 16:58, said:

 

 

Hi guys!

 

For anyone who is having lag / server / ping issues - please do send us a Support ticket. We are aware of the situation and are collecting specific information to help diagnose and hopefully address this issue.

 

For the Support ticket, we need:

  • PingPlotter run while issues are occurring
  • WG Check while issues are occurring
  • Replay showing the lag

 

You can go here to find a nice guide on what is needed to report the issues.

 

All three items should be attached to the Support ticket. If there are any questions, please let me know.

 

It is calling time for sever upgrade  chuck out the gold crapput new crap in

Edited by ImABadTealREEEE, May 13 2017 - 19:15.


TeamTerrible #247 Posted May 13 2017 - 19:19

    Sergeant

  • Players
  • 4390 battles
  • 156
  • [S-S-G] S-S-G
  • Member since:
    12-29-2012

I love how WG's CEO stated he was personally overseeing the restructure of Tanks.....

 

As we can see, problems do start from the top :teethhappy:



badmonkey59 #248 Posted May 13 2017 - 19:29

    Captain

  • Players
  • 44017 battles
  • 1,414
  • [-_W_-] -_W_-
  • Member since:
    05-15-2011

View Postrich73, on May 13 2017 - 11:42, said:

 

You know this is just a traffic related event and there is nothing anyone can do about traffic.With more and more people streaming nowadays,the infrastructure is stressed and only upgrades in the area where the bottlenecks are, will fix anything.I myself was rerouted for almost 3 years to west server,resulting in a 130 ping versus a 20 to 30 ping I enjoy now.I had to play east at this time,and I live in the west, as the ping was better,funny thing being that east never changed the whole period but did get better when I was grandfathered to a better internet plan.The better internet plan had no effect on west.So internet traffic is affected by routing and this changes as there is more traffic and switching comes into effect,resulting in being routed a farther distance,which gives worse performance.So rather blame streaming and insufficient upgrading of the infrastructure,tho if W G can improve their performance somehow to maybe compensate somehow,that would be welcome too.

 

This sounds good, but the lag problem started for thousands of players with the introduction of 9.17. To be clear, no problem until the update and then a problem. No problems with ANYTHING else. Wargaming has finally admitted there is a problem, they gave everyone 3 days of premium because of it.

Black_Stealth_Badger41 #249 Posted May 15 2017 - 02:20

    Captain

  • Players
  • 39300 battles
  • 1,749
  • [DBN] DBN
  • Member since:
    07-17-2012

View Postdance210, on May 13 2017 - 07:58, said:

 

 

Hi guys!

 

For anyone who is having lag / server / ping issues - please do send us a Support ticket. We are aware of the situation and are collecting specific information to help diagnose and hopefully address this issue.

 

For the Support ticket, we need:

  • PingPlotter run while issues are occurring
  • WG Check while issues are occurring
  • Replay showing the lag

 

You can go here to find a nice guide on what is needed to report the issues.

 

All three items should be attached to the Support ticket. If there are any questions, please let me know.

 

When you play for 4 hours and have 600+ packet drop over that time Ping Spikes at intermittent intervals. Everything off the Ping Plotter and in connection support through WTFast point directly to your servers..hmmmmm

HeavyMetalDroid #250 Posted May 15 2017 - 04:28

    Corporal

  • Players
  • 23480 battles
  • 22
  • Member since:
    01-13-2013
My computer and ISP don't lag, your game and servers lag.

KupJones #251 Posted May 15 2017 - 15:17

    Corporal

  • -Players-
  • 35127 battles
  • 47
  • [SYTHE] SYTHE
  • Member since:
    06-29-2014

View Postrich73, on May 13 2017 - 11:42, said:

 

You know this is just a traffic related event and there is nothing anyone can do about traffic.With more and more people streaming nowadays,the infrastructure is stressed and only upgrades in the area where the bottlenecks are, will fix anything.I myself was rerouted for almost 3 years to west server,resulting in a 130 ping versus a 20 to 30 ping I enjoy now.I had to play east at this time,and I live in the west, as the ping was better,funny thing being that east never changed the whole period but did get better when I was grandfathered to a better internet plan.The better internet plan had no effect on west.So internet traffic is affected by routing and this changes as there is more traffic and switching comes into effect,resulting in being routed a farther distance,which gives worse performance.So rather blame streaming and insufficient upgrading of the infrastructure,tho if W G can improve their performance somehow to maybe compensate somehow,that would be welcome too.

Its a bit more complicated than that.  I'm going to make some assumptions here, largely based on what I have done with other customers.  And I am only looking at NA EAST servers

 

ASSUMING:

WG hosts their server farm with a hosting firm.  That hosting firm is basically at a peering point -- ie., multiple ISP's converge on the hosting fiirm's facility.  Because of this arrangement, one's traffic can ingress the WG farm via more than one route -- that route determined by *your* ISP and how your ISP routes.

 

Take me -- I have 2 Internet connections.  Both ISP's peer into different backbone providers -- One, Cogent the other AS6453.net (which is TATA comms).  Latency on the Cogent side runs 80-90ms, starting with path 9.17.1.  Latency on the AS6453 side, however, still runs 50ms.  Now, one would immediately think "Easy.  Cogent is traffic buried while AS6453 is not!"  Au contraire -- BOTH backbones are running (from my perspective) at about 50ms latency with 0 packet loss in the core  -- which tells us that both backbone providers are handling the traffic pretty well.

 

Make Note Of This Address:  92.223.118.164 -- that is owned by CORE.PW, which is Core-Labs.

 

Then things change when we exit the backbone.   Coming off of Cogent we hit a device that is addressed out of the 92.223.x.x address space and which sits RIGHT IN FRONT OF the WG server load balancer.  This device is typically 92.223.118.170, but has been .163 before.  THIS DEVICE introduces about 30ms of latency, and only started doing so at patch 9.17.1 (I believe).

Now we get really interesting, because coming off of the AS6453 core we completely bypass this

 

Now, here is the odd part.  Traffic coming from AS6453 completely bypasses 92.223.118.170  and hits the WOT load balancer.  Result: 40-50ms latency from end to end.   This simple test is the clearest indication that, from where I sit, the issue is not the target server(s).

 

That single device (92.223.118.170) that is introducing the latency is owned by Core-Labs --- and it so happens that the 92.223 address space is also used to access the RU and EU server clusters.  But one only hits this device if one comes off of the Cogent Communications backbone.   So, COGENT ----> CORE-Labs ----> WOT servers.   And the Core-Labs device is the LAST HOP before hitting the WOT servers, telling me that this is one of the companies that is peering at the server hosting facility -- and therefore contractually answerable to WarGaming.

 

Of course, I could be totally wrong in my analysis.  But in the meantime I have reconfigured my router to prefer the path through AS6453 and I once again have 40-50ms latency.   Does anyone need more proof about where the problem is located?

 

PS, I have opened many tickets, have provided many traces -- and have received little back other than "we are looking into it" or "you need to see if it is your provider that is the issue."   The easiest way to make a problem go away is to simply pretend that one doesn't have a problem.

 

PPS.  The NA WEST servers are simply in a pickle.  The AS6453 backbone is throwing away packets to the tune of 20-30% and WG cannot fix that EXCEPT to move to a hosting provider who peers through another backbone provider.



Cognitive_Dissonance #252 Posted May 15 2017 - 18:10

    Major

  • Players
  • 40641 battles
  • 6,402
  • [ANASS] ANASS
  • Member since:
    01-31-2013

View Postkupjones, on May 15 2017 - 08:17, said:

Its a bit more complicated than that.  I'm going to make some assumptions here, largely based on what I have done with other customers.  And I am only looking at NA EAST servers

 

ASSUMING:

WG hosts their server farm with a hosting firm.  That hosting firm is basically at a peering point -- ie., multiple ISP's converge on the hosting fiirm's facility.  Because of this arrangement, one's traffic can ingress the WG farm via more than one route -- that route determined by *your* ISP and how your ISP routes.

 

Take me -- I have 2 Internet connections.  Both ISP's peer into different backbone providers -- One, Cogent the other AS6453.net (which is TATA comms).  Latency on the Cogent side runs 80-90ms, starting with path 9.17.1.  Latency on the AS6453 side, however, still runs 50ms.  Now, one would immediately think "Easy.  Cogent is traffic buried while AS6453 is not!"  Au contraire -- BOTH backbones are running (from my perspective) at about 50ms latency with 0 packet loss in the core  -- which tells us that both backbone providers are handling the traffic pretty well.

 

Make Note Of This Address:  92.223.118.164 -- that is owned by CORE.PW, which is Core-Labs.

 

Then things change when we exit the backbone.   Coming off of Cogent we hit a device that is addressed out of the 92.223.x.x address space and which sits RIGHT IN FRONT OF the WG server load balancer.  This device is typically 92.223.118.170, but has been .163 before.  THIS DEVICE introduces about 30ms of latency, and only started doing so at patch 9.17.1 (I believe).

Now we get really interesting, because coming off of the AS6453 core we completely bypass this

 

Now, here is the odd part.  Traffic coming from AS6453 completely bypasses 92.223.118.170  and hits the WOT load balancer.  Result: 40-50ms latency from end to end.   This simple test is the clearest indication that, from where I sit, the issue is not the target server(s).

 

That single device (92.223.118.170) that is introducing the latency is owned by Core-Labs --- and it so happens that the 92.223 address space is also used to access the RU and EU server clusters.  But one only hits this device if one comes off of the Cogent Communications backbone.   So, COGENT ----> CORE-Labs ----> WOT servers.   And the Core-Labs device is the LAST HOP before hitting the WOT servers, telling me that this is one of the companies that is peering at the server hosting facility -- and therefore contractually answerable to WarGaming.

 

Of course, I could be totally wrong in my analysis.  But in the meantime I have reconfigured my router to prefer the path through AS6453 and I once again have 40-50ms latency.   Does anyone need more proof about where the problem is located?

 

PS, I have opened many tickets, have provided many traces -- and have received little back other than "we are looking into it" or "you need to see if it is your provider that is the issue."   The easiest way to make a problem go away is to simply pretend that one doesn't have a problem.

 

PPS.  The NA WEST servers are simply in a pickle.  The AS6453 backbone is throwing away packets to the tune of 20-30% and WG cannot fix that EXCEPT to move to a hosting provider who peers through another backbone provider.

 

AS (Autonomous System) 6453 is a BGP (Border Gateway Protocol - Backbone carrier routing protocol) peering node. It is not so much a device, as a backbone carrier router(s) hosting about 175K IPV4 supernets/route summarization, or pointers to large blocks of public nets. It would not be easy to move to a different provider or hosting provider in the sense I think you are using it. It would be difficult for people to "use a different one" as this is a central node hosting lookup's to major network CIDR/Variable Length Subnet Mask boundaries.

 

BGP backbone routers are assigned ASN's (Autonomous System Numbers, i.e. 6453) that they use to establish comms with other large BGP federations/communities about where large blocks of IPV4/IPV6 addresses reside. This may be the only feasible route, based on "cost", even though other routes exist, and how this ASN peers to other networks.


Edited by SmirkingGerbil, May 15 2017 - 18:13.


KupJones #253 Posted May 15 2017 - 18:25

    Corporal

  • -Players-
  • 35127 battles
  • 47
  • [SYTHE] SYTHE
  • Member since:
    06-29-2014

View PostSmirkingGerbil, on May 15 2017 - 11:10, said:

 

AS (Autonomous System) 6453 is a BGP (Border Gateway Protocol - Backbone carrier routing protocol) peering node. It is not so much a device, as a backbone carrier router(s) hosting about 175K IPV4 supernets/route summarization, or pointers to large blocks of public nets. It would not be easy to move to a different provider or hosting provider in the sense I think you are using it. It would be difficult for people to "use a different one" as this is a central node hosting lookup's to major network CIDR/Variable Length Subnet Mask boundaries.

 

BGP backbone routers are assigned ASN's (Autonomous System Numbers, i.e. 6453) that they use to establish comms with other large BGP federations/communities about where large blocks of IPV4/IPV6 addresses reside. This may be the only feasible route, based on "cost", even though other routes exist, and how this ASN peers to other networks.

 

True, in the original use of the term Autonomous System and as it applies to how routing is performed - and listing out BGP routing on each router will provide an AS "hop" listing, showing which AS' a packet has traversed.

 

HOWEVER, AS6543.net is a DNS domain created by TATA and traces indicate how packets transit across the TATA backbone - so they have subverted the true meaning of the term "AS".   So my statement above still hold.

 

Plus, lets add this.  Tracing from an entirely different location (my employer) I transit yet another backbone provider and latencies stay within 60ms UNTIL they hit 92.223.118.170, then 30ms of latency is added by that device.

 

Folks, this stuff isnt hard -- it just takes some sleuthing to build the picture.  Someone from WG really needs to get in touch with Core-Labs.


Edited by kupjones, May 15 2017 - 20:40.


Cognitive_Dissonance #254 Posted May 15 2017 - 21:32

    Major

  • Players
  • 40641 battles
  • 6,402
  • [ANASS] ANASS
  • Member since:
    01-31-2013

View Postkupjones, on May 15 2017 - 11:25, said:

 

True, in the original use of the term Autonomous System and as it applies to how routing is performed - and listing out BGP routing on each router will provide an AS "hop" listing, showing which AS' a packet has traversed.

 

HOWEVER, AS6543.net is a DNS domain created by TATA and traces indicate how packets transit across the TATA backbone - so they have subverted the true meaning of the term "AS".   So my statement above still hold.

 

Plus, lets add this.  Tracing from an entirely different location (my employer) I transit yet another backbone provider and latencies stay within 60ms UNTIL they hit 92.223.118.170, then 30ms of latency is added by that device.

 

Folks, this stuff isnt hard -- it just takes some sleuthing to build the picture.  Someone from WG really needs to get in touch with Core-Labs.

 

No, they have not, it is a true Autonomous System, registered and holding (currently) 175K IPV4 supernets/prefixes, and over 7K IPV6 prefixes. Of course they could be doing DNS names with it, but if you use tracert or any other tracing tool, if a transit hop allows trace route to do name lookups, you will get a name like ia.brdr.asn65000.chi.go as an example.

 

It is hard, it is what I do for my job, seeing as how I maintain internal private ASN's feeding off of ATT BGP routers, I can say, it isn't as simple as you paint it. They are the transit owner for 175,000 V4 prefixes, it is registered with ARIN (American Registry of Internet Numbers) as an AS, seeing as how ARIN hands out ASN's and that number is registered to TATA, pretty sure it is not "subverted".

 

 



badmonkey59 #255 Posted May 16 2017 - 22:13

    Captain

  • Players
  • 44017 battles
  • 1,414
  • [-_W_-] -_W_-
  • Member since:
    05-15-2011

View PostSmirkingGerbil, on May 15 2017 - 14:32, said:

 

No, they have not, it is a true Autonomous System, registered and holding (currently) 175K IPV4 supernets/prefixes, and over 7K IPV6 prefixes. Of course they could be doing DNS names with it, but if you use tracert or any other tracing tool, if a transit hop allows trace route to do name lookups, you will get a name like ia.brdr.asn65000.chi.go as an example.

 

It is hard, it is what I do for my job, seeing as how I maintain internal private ASN's feeding off of ATT BGP routers, I can say, it isn't as simple as you paint it. They are the transit owner for 175,000 V4 prefixes, it is registered with ARIN (American Registry of Internet Numbers) as an AS, seeing as how ARIN hands out ASN's and that number is registered to TATA, pretty sure it is not "subverted".

 

 

 

How about you send WG an email and tell them how to fix it. No one there seems to be able to figure it out.

Cognitive_Dissonance #256 Posted May 17 2017 - 14:57

    Major

  • Players
  • 40641 battles
  • 6,402
  • [ANASS] ANASS
  • Member since:
    01-31-2013

View Postbadmonkey59, on May 16 2017 - 15:13, said:

 

How about you send WG an email and tell them how to fix it. No one there seems to be able to figure it out.

 

Lol, honestly, I wish I could, but to be realistic it probably isn't a quick fix and would require a serious overhaul in who is peering to them (read new contracts and migrations.)

 

I am migrating an IP backbone now, from legacy to newer ATT tech, and we just keep hitting one snag after another.



wilwoods #257 Posted May 30 2017 - 00:53

    Sergeant

  • Players
  • 31272 battles
  • 208
  • Member since:
    08-22-2013

Dance I am in Toronto area and the lag is still pretty bad. I usually run about 100 fps and 22 ms. Now Its falling to 5 fps and 500 ms. The fps has climbed to as high as 70 and ping is sometimes 100 and it has got back down to 22 but right now, or at least for the last 2 weeks its been around 85.

Tonight May 29 its 35 fps and 85 ms.

 

The numbers really started dropping after the last patch. I do use xvm but mini map is minimized. I have an ethernet connection to my router and  fast PC with 32 GB of RAM

 



Lowsmoke #258 Posted May 30 2017 - 14:52

    Corporal

  • Players
  • 19003 battles
  • 73
  • Member since:
    12-30-2011
Lag is caused by people with high pings in game dont believe their bs . If  they had another server in brazill which they willnt do cause they have no money to milk out of them what a bunch a crap these dudes try to give u all.... They r russians lol they lie .F  ur nerf and ur stupid rules.Ban me  do me a favor ..Urs truly a real gamer Lowsmoke  game blows

Kliphie #259 Posted May 30 2017 - 21:09

    Major

  • Players
  • 32828 battles
  • 5,363
  • [GFLC] GFLC
  • Member since:
    07-20-2012

View PostLowsmoke, on May 30 2017 - 08:52, said:

Lag is caused by people with high pings in game dont believe their bs . If  they had another server in brazill which they willnt do cause they have no money to milk out of them what a bunch a crap these dudes try to give u all.... They r russians lol they lie .F  ur nerf and ur stupid rules.Ban me  do me a favor ..Urs truly a real gamer Lowsmoke  game blows

 

  The other 29 players have no effect on your lag, that's not how a server based game works.  

KupJones #260 Posted May 31 2017 - 01:21

    Corporal

  • -Players-
  • 35127 battles
  • 47
  • [SYTHE] SYTHE
  • Member since:
    06-29-2014

View PostSmirkingGerbil, on May 15 2017 - 14:32, said:

 

No, they have not, it is a true Autonomous System, registered and holding (currently) 175K IPV4 supernets/prefixes, and over 7K IPV6 prefixes. Of course they could be doing DNS names with it, but if you use tracert or any other tracing tool, if a transit hop allows trace route to do name lookups, you will get a name like ia.brdr.asn65000.chi.go as an example.

 

It is hard, it is what I do for my job, seeing as how I maintain internal private ASN's feeding off of ATT BGP routers, I can say, it isn't as simple as you paint it. They are the transit owner for 175,000 V4 prefixes, it is registered with ARIN (American Registry of Internet Numbers) as an AS, seeing as how ARIN hands out ASN's and that number is registered to TATA, pretty sure it is not "subverted".

 

 

 

Just saw this -- and dammed if you aren't right.  I wasnt aware that a a referral could be made to find an AS and return it as a DNS-like name.  Kinda cute - definitely threw me off.  See, we all learn something everyday -- and I do networking for a living BUT not as an SP like you do.   

 

I did manage to re-route my traffic via another SP (Centurylink) and appear to be avoiding the 92.223 IP block.  This has the result of not running through that provider and ping times have fallen to 35ms.  Still does nto change the fact that they have a serious issue with their peering arrangements at their co-lo.

 

Thanks Smirking!  






1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users