Dropped Packets between my PC and game server

Original post edited for updated data.

I'm getting a lot of dropped packets in between my network and my game server (Figaro). Everything looks good on my end up until it gets to SE's internal network. Could you look into your hardware to see if you can optimize your network performance? This post is a week old now and this problem has only occurred for me since patch 1.19a. If you look in some of the other threads in the tech support forum, other people are having the same issues.

I've provided some ping and traceroute results. The following tests were performed starting 8:45pm on 10/19/2011 pacific time. Windows task manager showed that my network utilization was only 0.02% as these tests were being performed. Sorry for the spam:

Tracing route to 202.67.48.202 over a maximum of 30 hops

1 9 ms 9 ms 7 ms 10.145.160.1
2 8 ms 8 ms 7 ms 68.2.8.45
3 10 ms 9 ms 11 ms 70.169.75.86
4 11 ms 5 ms 4 ms mcdldsrj01-ae2.0.rd.ph.cox.net [70.169.76.225]
5 20 ms 20 ms 21 ms langbprj01-ae1.rd.la.cox.net [68.1.1.13]
6 137 ms 137 ms 138 ms gi6-0-0.cr1.nrt1.asianetcom.net [202.147.61.166]
7 137 ms 137 ms 144 ms gi1-0-0.gw1.nrt5.asianetcom.net [202.147.0.178]
8 164 ms 136 ms 137 ms squareco.asianetcom.net [203.192.149.210]
9 148 ms 139 ms * 61.195.56.129
10 155 ms 155 ms 156 ms 219.117.144.2
11 154 ms 156 ms 155 ms 219.117.146.149
12 147 ms 148 ms 147 ms 219.117.146.129
13 147 ms 138 ms 138 ms 219.117.146.14
14 134 ms 155 ms 145 ms 202.67.48.202

I get no dropped packets up until Hop #9 which is the first router past SE's edge router. Hops 9-11 all have dropped packets. Hops 12 and 13 strangely have a 0% drop rate. Hop 14 is the server and drops packets. Ping results are below:

Ping statistics for 61.195.56.129:
Packets: Sent = 1000, Received = 822, Lost = 178 (17% loss),
Approximate round trip times in milli-seconds:
Minimum = 130ms, Maximum = 215ms, Average = 148ms

Ping statistics for 219.117.144.2:
Packets: Sent = 1000, Received = 826, Lost = 174 (17% loss),
Approximate round trip times in milli-seconds:
Minimum = 131ms, Maximum = 179ms, Average = 147ms

Ping statistics for 219.117.146.149:
Packets: Sent = 1000, Received = 838, Lost = 162 (16% loss),
Approximate round trip times in milli-seconds:
Minimum = 129ms, Maximum = 180ms, Average = 147ms

Ping statistics for 219.117.146.129:
Packets: Sent = 1000, Received = 1000, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 141ms, Maximum = 185ms, Average = 148ms

Ping statistics for 219.117.146.14:
Packets: Sent = 1000, Received = 999, Lost = 1 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 134ms, Maximum = 183ms, Average = 141ms

Ping statistics for 202.67.48.202:
Packets: Sent = 1000, Received = 828, Lost = 172 (17% loss),
Approximate round trip times in milli-seconds:
Minimum = 132ms, Maximum = 180ms, Average = 150ms

This problem has gone on for a week for me now and other are having the same problem. PLEASE forward this onto your server/network technicians so they can get this sorted out. I want to be one of your paying customers once subscription starts up but if this can't get fixed what point is there me in playing if I have 5 seconds to 5 minutes of lag in between each of my actions?

Just wanted to mention also, I spoke with a Tier 2 technician at my ISP (Cox Communications) and because there are no dropped packets until I get inside SE's network and my ping times are really good, they are certain this is not a problem on their end.

I also spoke with a level 1 SE representative this morning and I have to say the interaction wasn't pleasant. He was quite condescending towards me as I tried to explain my problem and present evidence that it wasn't simply an issue of a router being hooked up to my network. He claims he cannot have proper testing performed with a router hooked up, but I have not had ANY connectivity issues prior to patch 1.19a, but he wouldn't hear any of it.

Bumping this thread because it needs to be brought to the attention of server/network technicians at SE. The 30002 error is the result of the game lagging out due to the connection dropping for too long of a period.

Hope we can get this issue solved Oro

Ya this has been determined by us the players already. There's a couple other threads about this really. I'm hoping they discover what is going on soonish.

Edit: I didn't realize you had this open for awhile my bad lol.

This is one of the reasons theire making a new engine. In 2.0 it should be fixed, hopefully.

Why is is everytime someone starts a thread about some technical issue there's always someone that comes in and is like HEY 2.0 will fix it don't worry. Its kinda getting old. Actually the correct answer is they are making new servers to fix the problems related to this, and the slow UI. I will submit however if the engine is designed to process many things on the client side instead of on the server side, then that could also go a long way to help.

Quote:

---

Originally Posted by Medura

This is one of the reasons theire making a new engine. In 2.0 it should be fixed, hopefully.

---

This is not a problem of the game engine. It's the company's general firewall.

Anyway, don't rely on ping or nslookup here completely, even though it looks suspicious at first sight. It's ICMP, a completely different protocol on a different network layer, and it's like comparing apples and oranges. I'm quite surprised that the Square routers reply to ICMP at all, which is quite uncommon for companies of this sizes ... but well, it's a company of developers and not administrators and this had been probably outsourced to the guy with the nicest powerpoint, too.

Orophin, the best thing to do would talk to a friend of yours who uses a different ISP than you do and when you observe this again, your friend should check exactly the same IPs on his side, too. Best do this synchroneously when talking on the phone and using a different ISP is of importance to make sure it's not a routing issue which only affects certain address ranges, resulting in CRCs hitting your packets or whatever. If he can not reproduce the problem, then it's probably not directly related to SE (only indirectly).

fyi, I could not reproduce the problem on my side, I had no packet losses. And even if, it's still TCP with flow control and retransmits (which ICMP and thus your ping hasn't).

If your friend can reproduce the problem, then download wireshark (http://www.wireshark.org/) and create a protocol with your ICMP traffic (ping & traceroute for 30-60 seconds) and another one when you are having your symptons while playing the game (you can start wireshark after the problems arise, no need to log off). Make sure you capture the traffic to all the SE networks, so 61.195.* and 219.117.* should suffice as long as you have no other parallel traffic to Tokyo areas. If all that networking stuff is not really clear, then you could ask your ISP to monitor your packets directly on the first switch or router ... or even better, directly ping from their router and check if they can ackonwledge your findings - they can do that, even over longer periods of time, it's part of the daily business of any network administrator and it's just a question if they want to and/or if your support contract with your ISP covers this.

Also make sure that your router is not the problem. Ask a friend to lend him yours, or whatever. Then try to reproduce the problem with his router, too. If the problem disappears, then there may be the root cause (packet filter, cyclic redundancy checks, or perhaps packet inspections, some of todays cheap router try to act smarter than they actually are) or whatever.

When you can reproduce the problem with different router hardware AND a different ISP, then this needs to be adressed directly to SE networking operations. the difficulty will be to find out how. Don't let yourself stressed being out by a first level supporter of the game, that's a) the wrong recipient because this has nothing to do with the game itself and is outside of his responsibilities, and b) he or she probably does not really understand what the problem is. I would directly demand an escalation to the networking team - but when you do that, it would be vital to deliver them a wireshark capture (or, if your ISP can do this for you, a tcpdump / snoop / IOS related ping results / whatever).

Hope that helps a bit to get the train back on the tracks.

@Medura:

This is something they broke with the current patch, not some ongoing problem a lot of people have been having for a while. If you can't contribute to this thread, please don't post. We're looking for a resolution here.

Actually they keep fixing and breaking it over and over. I think SE just has a general problem with keeping their hardware alive... Can just hope they will take better care when they have paying customers :/

bump
bump bump bump bump

SolArisa:

Thanks for your input. I ran a wireshark capture tonight and posted my findings in a spreadsheet.

I have a spreadsheet I've uploaded online for review that hopefully will give some insight into what this problem is. Here's a summary of what's going on:

* 22 frames had a duplicate acknowledgement to a frame before it due to the a segment in the frame before it being lost.

* 62 frames had a tag of [TCP Previous segment lost].

* EVERY packet sent FROM the server with the exception of those listed above (items 1 and 2) was received with no errors.

Wireshark link: https://www.filesanywhere.com/fs/v.a...8a5e626ebd73ab

I can upload the actual wireshark log file if network support needs it. It's just a proprietary format and won't really work without wireshark I don't think.

Now the question is, what causes this problem? I got back on the phone with Cox level 2 support but he wasn't really able to help with something this technical so he gave me an e-mail address to send my data to. I'm hoping they can decipher something out of it that makes sense.

Uploaded a new file tonight:

http://www.filesanywhere.com/fs/v.as...8c5c676da76e9e

The packets in question have the following tags:

[TCP Dup ACK]
[TCP Retransmission]
[TCP Previous segment lost]

The first tag I've read can be a problem due to a duplex mismatch between network devices. By default, duplex is set to auto, which my NIC is. I've e-mailed my ISP just to ensure that nothing changed on their end, but it seems highly unlikely something changed on their end the day after patch 1.19a hit. Perhaps if someone from SE would actually take the time to read these threads, they can check either:

a) their coding for the latest patch to make sure packets aren't being sent in something other than Auto Duplex
b) their hardware for any duplex changes.

Support for this game is absolutely terrible. This is where we get directed to when we need an issue fixed. And the only thing that ever gets posted is them asking for your PC specs. If they start charging for this game next month I will not be playing it that is for certain.. Its sad when you have people that want to play the game and can't....

Quote:

---

Originally Posted by Fraus

Support for this game is absolutely terrible. This is where we get directed to when we need an issue fixed. And the only thing that ever gets posted is them asking for your PC specs. If they start charging for this game next month I will not be playing it that is for certain.. Its sad when you have people that want to play the game and can't....

---

I completely agree. This is a typical Square Enix response. If they would only acknowledge that they recognize that there is even a problem, then that would be something.

The basic problem is that this ONLY STARTED HAPPENNING after patch 1.19a. It is being reported by many people in differtne locations with different ISPs. Therefore, it is NOT an ISP issue. It is an issue with SE hardware.

For any of them to even slightly imply that it could be our hardware is ridiculous. They developed the benchmark program. We passed the benchmark. Many of us built PCs just in order to play FFXIV.

They need to address this issue and get in-game GMs that you can personally complain to about RMT and botting activity and have the authority to take immediate action. (I'm not attempting to derail, but it is typical of their customer support)

I have posted on other 30002 threads, but i can 100% say its not any problem on our (players) end, i can log in an alternate character on same account, with no issues, but my main character, gets 30002 at now loading screen when i attempt to log in, if customer service does not fix this TODAY, I will be cancelling my FFXI subscription, and will not be playing either game anymore, pretty ridiculous imo...

I definitely think you should blood I wouldn't give them any money right now. When as many people have connection issues as they do that should be first priority.. I even completely reinstalled and downloaded the patch thinking it would help and it wouldnt... And they don't even post on here or make an announcement acknowledging they are aware of our issues.