375 replies to this topic

[Tuku]

Thanks

[Tuku]

Intresting My Ping and Trace routes all show what my normal average ping is with one request timed out otherwise its normal....yet im still having this problem in game.

[NocturnalBeast]

I am honestly getting tired of troubleshooting this. I am not getting paid to be a PGI, or TWC network admin and PingZapper works more often than not, so I will just live with it and hope that the people who are getting paid to solve this, solve this.

[Jon Cunningham]

Ed Steele, on 07 February 2015 - 03:29 AM, said:

Ok, still needed to use Pingzapper tonight to get sub 200ms pings. I did get a couple disconnects though using the US Central North servers. US West VPN servers are guaranteed disconnect at this time. Jon have you contacted Pingzapper yet to work on that profile?

Yep, both them and Wtfast.

Pingzapper is working with me now to add official support.

Ed Steele, on 09 February 2015 - 05:58 PM, said:

I am honestly getting tired of troubleshooting this. I am not getting paid to be a PGI, or TWC network admin and PingZapper works more often than not, so I will just live with it and hope that the people who are getting paid to solve this, solve this.

Good to hear Pingzapper is working.

We're still working on getting TWC to play nice, but until they do my only option is to start reaching out to VPN providers and getting them to add official support for our game.

Please enjoy!

[NocturnalBeast]

Jon Cunningham, on 10 February 2015 - 11:47 AM, said:

Yep, both them and Wtfast.

Pingzapper is working with me now to add official support.




Good to hear Pingzapper is working.

We're still working on getting TWC to play nice, but until they do my only option is to start reaching out to VPN providers and getting them to add official support for our game.

Please enjoy!

Your effort is appreciated.

[Clit Beastwood]

My recent interactions with TWC have been great - they came on-site, replaced a bunch of cabling etc (eliminating variables) - since that fixed nothing, we're progressing through the support levels. Progress!

[LordSkippy]

I called TWC and had them put my modem in bridge mode without the port blocking. Didn't make too much of a difference, maybe 3-5ms less. So, instead of the mid 80s, it's the low 80s. Haven't had any huge ping spikes either. Pingzapper is now only shaving 5-10ms off the ping, so I've pretty much stopped using it. Haven't noticed a difference with hit detection either way.

[Santanico Pandamonium]

I have previous posted that I have been having the same issue as everyone else, being from Central Texas and Time Warner.

However, after the last patch, I've not used pingzapper. And for almost 2 weeks, not experienced any "ping" issues. My ping isn't as high as it once was (40ms) but it is a constant 80-90 through the night, to include prime time.

I don't know if there has been any optimizatons or what not on PGI's part or the ISPs involved, but my problems are gone for now.

[LordSkippy]

Personally, I think the throttling stop when an accounting period cycled over. Putting TWC back below their contracted bandwidth throughput with a peer, thus ending the throttling.

Of course, that's just my personal conspiracy theory.

[Jon Cunningham]

Hey folks,

Just an update. It looks like for quite a few users this issue has been abating. We have been working out alternate paths, working with our DC partner(s) to acquire new peers, optimizing our traffic with as many regional entities as we can manage, and working with VPN providers to get around throttling issues.

I want to thank everyone in the community that has contributed information, contacts, debug steps, for folks who have opened tickets with TWC or put pressure on them. Especially Fierostetz, Kin3ticX, SoggyGorilla, Fierostatz and others.

While optimizing this will still be an ongoing issue, especially for folks getting hit by regional throttling, a lot of progress has been made since we started working on it.

[Clit Beastwood]

Thanks Jon! My huge lag spikes have cleared up over the last several days - hitreg is still pretty iffy, but I'm not dropping 80% of my packets in transit anymore. No rubberbanding, ballistics work fine, ppc's are iffy, and lasers seem to do about 75% which is a drastic improvement.

[Kin3ticX]

Jon Cunningham, on 13 February 2015 - 11:16 AM, said:

Hey folks,

Just an update. It looks like for quite a few users this issue has been abating. We have been working out alternate paths, working with our DC partner(s) to acquire new peers, optimizing our traffic with as many regional entities as we can manage, and working with VPN providers to get around throttling issues.

I want to thank everyone in the community that has contributed information, contacts, debug steps, for folks who have opened tickets with TWC or put pressure on them. Especially Fierostetz, Kin3ticX, SoggyGorilla, Fierostatz and others.

While optimizing this will still be an ongoing issue, especially for folks getting hit by regional throttling, a lot of progress has been made since we started working on it.

Yay! I have noticed an improvement. I have been running without a VPN for a few days. You should post this on hpg so that everyone there knows they should be on the lookout for improvements.

[aniviron]

I'm on TWC in Texas, and have noticed a significant improvement. My pings have gone from ~120 to ~80, and I did not have a single game spike to 500+ while completing the challenge.

[eta0h]

Fierostetz, on 10 December 2014 - 10:13 AM, said:

I'm well versed in network testing and diagnostics, if I'm able to be of any assistance I'll gladly grab logfiles, etc. I'm running a pfsense firewall with verbose logging. From what I've seen, at the beginning and ending of every drop there's a flurry of packet drops on my firewall - when I had a different OS running on my hardware router appliance it would detect a DOS attack whenever drops started or finished. I'm guessing the two are related - I'm wondering if TWC is filtering or throttling the data on MWO ports because, for all intents and purposes, it's unidentified traffic. If our routers are seeing this flurry of activity as a DOS attack, it's not outside the realm of possibility that TWC might be seeing it as the same thing. I can generally drop and get 2-3 drops in before my 92ms ping climbs to 400-500ms. These issues were not present prior to the server move. Related or not, I don't know, but my hitreg went out the window when the server move happened. Since then I've played far less MWO. Hitreg is noticeably better on my Dad's COX cable connection, and from my office's COX fiber connection.

Here's a snapshot of my traffic shaping queues. I cleared the queue stats right before playing for a little while last night. I played a few drops on my regular connection (the qgames queue) and 1-2 drops via VPN (tunneled and categorized into the "qdefault" queue (port 443)).


Here's how the traffic queues are defined - this is the MWO section rules


Here's the most recent traceroute (much improved from immedately after the server move, after the server move, for a while, I was getting 90 something hops to get out of SoCal)


Speedtest


Pingtest


Additional troubleshooting stuff I've also tried
- Switching between multiple Motorola SB6141 routers. (Best downstream performance)(I have a bunch of them).
- Using the new ARRIS router provided by TWC. (10+ minutes to sync?? WTF channel bonding!? Best upstream performance).
- Yes I've run it with the computer direct-connected to the modem (sb6141 and the new arris, model number not available at the moment) and there were no changes.

The lag spikes, etc. were not present in any other games or applications I've tried running, and I've been leaning hard on the network while testing some workstation virtualization stuff from work to home. No packet loss while doing that or anything else, but once I start the MWO client the packet drops start as soon as I click "launch".

In my case I have the Trendmicro DPI engine running on my home router. It reports my UDP game traffic as eMule, which is P2P traffic. I wonder if ISPs arn't throwing P2P (MWO packets) into a lower priority queue, which manifests as lag spikes in game when the ISPs buffers start to tail drop during busy periods?

Edited by eta0h, 14 February 2015 - 10:07 AM.

[Mogney]

my isp is charter, my PING is in the 40s, my FPS between 60-120. For the last few days I have been teleporting all over the map constantly. Is this possibly a result of the changes to fix the time warner issue?

[Clit Beastwood]

Mogney, on 14 February 2015 - 01:52 PM, said:

my isp is charter, my PING is in the 40s, my FPS between 60-120. For the last few days I have been teleporting all over the map constantly. Is this possibly a result of the changes to fix the time warner issue?

Unlikely - try doing a traceroute and look for any excessively bad hops, or get the free version of pingplotter and hit the mwo server.

eta0h, on 14 February 2015 - 09:55 AM, said:

In my case I have the Trendmicro DPI engine running on my home router. It reports my UDP game traffic as eMule, which is P2P traffic. I wonder if ISPs arn't throwing P2P (MWO packets) into a lower priority queue, which manifests as lag spikes in game when the ISPs buffers start to tail drop during busy periods?

That was long my suspicion, but over the past few days it IS improving.

Edited by Fierostetz, 14 February 2015 - 04:30 PM.

[NocturnalBeast]

I have played without Pingzapper since Friday and my ping is staying between 100-150ms, which is good enough that I may not resub to Pingzapper, although it is still higher than before the server move.

[Clit Beastwood]

Well at this point I'd say that the issue is mostly resolved. I still get slightly better hitreg when playing from elsewhere, but the difference between a month ago and now is nothing short of miraculous. TWC did some stuff on my end - the most recent guy to come out replaced a good section of cable and redid the ground, etc. - I upgraded to an sb6183 modem (that was finally able to sync after they re-cabled stuff). That alone didn't make a big difference, dropped the ping a bit, but a day or two after I noticed far less rubberbanding and more ping stability. I'm attaching some graphs from the firewall because you can tell the *minute* that I was up and running on the new sb6183 modem with the new (not 25 year old) cabling.




As well as a pingplotter from this evening, no VPN


As well as with VPN (right now)


Thanks PGI guys for your efforts on our behalf - it's much appreciated, and it's even better to get validation that something we suspected was a problem... actually was!

[Dagorlad13]

FYI Pingzapper has a Mechwarrior Online profile now.

[Dauntless Blint]

Lag seems to be a bit of a problem since the release of VOIP,my 2 cents....and maybe the code going on for the Urbie messing with the F key centring function or what every it is.