Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 30 40 50 .. 52 :: one page |
|
Author |
Thread Statistics | Show CCP posts - 8 post(s) |
shifty one
|
Posted - 2009.03.23 01:36:00 -
[691]
from #3 petition that I got from a gm: Hi there,
It's quite normal to see packetloss at certain hops in Pingplotter. That is just happening because those servers are not responding to the requests made by Pingplotter. However, unless you also see packetloss at every hop after that, all the way to the last hop, then all the packets are reaching the final destination and the hop that reported the packetloss is not an issue. Therefore the packetloss reported at telecity-redbus-gw1.ip.tiscali.net is not relevant unless it were to continue reporting packetloss through the last hop. Please keep in mind that there are also thousands of people going through that hop and experiencing no problems.
We are still trying to find the source of this issue and that is being worked on with high priority. We ask for Pingplotter logs from players as we can often help players to prevent it from happening, at least until the real source of the problem is found. In the last petition about this we asked if you could send us a Pingplotter file using TCP settings instead of the standard ICMP settings. It would be good if you could send us that file so that we can analyze it using different settings in the program. I'm including again the instructions for that below:
1) In pinglotter go to the menu Edit - Options 2) Go to Packet tab and add the following settings: - Packet type: TCP Packets - Time to wait for ping replies: 3000 - Packet send delay: 40 - Packet size: 40 - ToS/DSCP header byte: 00000000 - Starting hop: Final hop only - TCP Port: Other: 26000 - Rest of the options leave at default 3) Under "Address to trace" put the following value in the box: 87.237.38.200 4) In the bottom left corner in the box " of times to trace" put the value Unlimited 5) In the bottom left corner in the box "Trace Intervals" change it to 2.5 seconds 6) In the bottom left corner in the box "Samples to include" change it to ALL. 7) Click on "Trace" button in the bottom left corner.
While running the test, please try and stay logged into EVE. Let the test run until you lose connection to the server.
When it's completed save it as the sample set ( pp2 ).
Once you have saved the file, you can attach it to the email you sent us. If this petition was created ingame you can attach filed by going to our website and clicking the link "My Petitions". You will be asked to log in then it will display all your petitions. Find this petition and at the bottom of the page will be a section to upload files.
Best regards, The EVE Online Customer Support Team
all this tells me is that they don't see a isp issue as a ccp issue. which is whats been put out already by a few people from this topic. someone mentioned they have contracts with the isp that has an issue with the hop(s). just an fyi- part of that contract would give a qos(quality of serive), I work for an isp- comcast. I deal with business accounts, and at least here in the states, there is a qos standard that has to be met, I dono if it's the same over in the UK, but I'm sure it is. and if thats the case, ccp could threaten to pull the contract if they didn't fix the issue. but coming full circle, this isn't a mass majority who are having the problem, it's a fraction. it comes back to the basics of squeaky wheel gets the grease, there isn't enough people complaining yet so they won't do anything. so all in all we're pretty much screwed. only thing you can do is get everyone with this issue to complain until it's fixed.
|
kurika
|
Posted - 2009.03.23 02:20:00 -
[692]
I'm also getting disconnected from eve. CCP please help us out. I did not have this problem until the new patch.
|
shifty one
|
Posted - 2009.03.23 02:38:00 -
[693]
I've had this issue on and off for the last few years.
|
Grall Sek
Caldari CREO Technologies
|
Posted - 2009.03.23 02:44:00 -
[694]
with those pingplotter settings i get error in first hop and it seems i cant get to the rest, does it mean i have some problem on my side? anyone knows? |
Carreira
|
Posted - 2009.03.23 10:01:00 -
[695]
Originally by: shifty one from #3 petition that I got from a gm: Hi there,
It's quite normal to see packetloss at certain hops in Pingplotter. That is just happening because those servers are not responding to the requests made by Pingplotter. However, unless you also see packetloss at every hop after that, all the way to the last hop, then all the packets are reaching the final destination and the hop that reported the packetloss is not an issue.
Oh this is wonderfull! I only have packet loss on one hop. Yet the GMs managing my petition never mentioned any of this, plus they specificaly claimed the issue is the 30% packet loss on the aforementioned Hop.
|
Venetian Tar
Violation of Freedom Kraftwerk.
|
Posted - 2009.03.23 10:14:00 -
[696]
Still getting this socket closed error, and it's really really annoying.
CCP not going to do anything about it?
I'm sure 24 pages would get your attention now...
|
Yashidza
|
Posted - 2009.03.23 13:20:00 -
[697]
before Apocrypha, had made petition and it helped, after Apocrypha disconnects started again 8( |
john roe
Children of Gjallarhorn
|
Posted - 2009.03.23 13:24:00 -
[698]
Originally by: Venetian Tar Still getting this socket closed error, and it's really really annoying.
CCP not going to do anything about it?
I'm sure 24 pages would get your attention now...
ORLY?
204 pages
here you have 204 pages (sic!) and as you can see, ccp dont give a **** about it. maybe when we'll reach 300 pages limit someone will do anything about it?
fingers crossed :]
peace
|
Lord Abbadon
Gallente Volition Cult
|
Posted - 2009.03.23 14:05:00 -
[699]
Get the same thing on both my computers, started since after expansion. Kinda annoying... |
Grall Sek
Caldari CREO Technologies
|
Posted - 2009.03.23 14:17:00 -
[700]
Ghost training whould be handy right now because im paying and i cant play the game, maybe if we get used to this socket closed issue and it becames a game feature they fix it |
|
Johren
|
Posted - 2009.03.23 15:43:00 -
[701]
Oddly enough, ping plotter shows much better results through that telecity hop after the expansion than before, but I'm still getting the occasional socket closed disconnects.
I got fed up with it Saturday night when I lost all 5 of my deployed Hammerhead II drones. I'm tired of having to buy those things back due to no fault of my own so this time I sent in a petition and asked for them to be replaced. Surprisingly the next day the replacement drones were sitting in my hangar. At least that went smoothly, though this would be much easier if the problem wasn't happening in the first place to all of us.
|
john roe
Children of Gjallarhorn
|
Posted - 2009.03.23 17:38:00 -
[702]
Originally by: Johren Oddly enough, ping plotter shows much better results through that telecity hop after the expansion than before, but I'm still getting the occasional socket closed disconnects.
I got fed up with it Saturday night when I lost all 5 of my deployed Hammerhead II drones. I'm tired of having to buy those things back due to no fault of my own so this time I sent in a petition and asked for them to be replaced. Surprisingly the next day the replacement drones were sitting in my hangar. At least that went smoothly, though this would be much easier if the problem wasn't happening in the first place to all of us.
ORLY?
i guess it only works for Hammerheads cause when i petition my n-th lost of Wasps t2... this is what i got instead :
"Hello.
I am sorry to hear about your loss. Unfortunately we are unable to reimburse this loss as our server-side logs do not indicate that a bug/error within the game or a server related problem was the reason for your loss. This does not mean that I doubt your description of events in any way but I'm afraid it keeps my hands tied in regards to reimbursement. Our reimbursement policies are very strict and I can only reimburse if I am able to verify that a bug or server error caused the loss. I hope that you understand our position and that you will recover swiftly from the loss.
Best regards,
GM GuruShura EVE Online Customer Support "
up yours, cccp o7
|
Aleksei Valenkov
|
Posted - 2009.03.24 08:57:00 -
[703]
I've been having this Socket Closed issue ever since Apocrypha. It happened sparsely at first then it got progressively worse. Today I've had over 10-20 disconnects in the span of an hour. It's ridiculous and it's starting to drive me nutty.
|
john roe
Children of Gjallarhorn
|
Posted - 2009.03.24 12:12:00 -
[704]
Originally by: Aleksei Valenkov I've been having this Socket Closed issue ever since Apocrypha. It happened sparsely at first then it got progressively worse. Today I've had over 10-20 disconnects in the span of an hour. It's ridiculous and it's starting to drive me nutty.
i know how to fix it... permanently :
suspend your accounts! till they fix it, prolly Soon(TM)
|
Ancient Darkstone
|
Posted - 2009.03.24 13:21:00 -
[705]
Originally by: Grall Sek with those pingplotter settings i get error in first hop and it seems i cant get to the rest, does it mean i have some problem on my side? anyone knows?
Not quite. First if you are on XP sp2+ or Vista, you will need Wincap libraries installed for PP to even be able to shape the tcp nop trace packets, and you will have to turn them on at the options editer - read the screen and it will tell you everything you need to know, as the ability to manipulate the tcp packets at this level was removed after sp2. Even so, that kind of packet manipulation may run afoul of any number of hops between you and the target. If you cant use icmp echo to connect there might be an issue.
Further, having run packet sniffers on the link, Eve will connect a talk on 26000 for a while with lots of failures until it finally just gets a packet transmission failure at the wrong point and hangs. The PP may still show you almost total failure on the echo trace settings even when you have partial comms.
The high packet loss should not of itself stop the game - just make it really cludgy and slow - as long as some packets get through every now and then. The problem looks very much like there are specific points in the application where the sending comms code just abandons a sent packet, or fails to re-request a critical packet that gets corrupted, or cant re-request data more than once, or that there are too many packets received after a failed packet and the packet queue buffer just runs out space waiting for a resent packet that was not received many packets ago, that it can not afford to drop (which is usually any data packet, as opposed to sound or graphics that can be fudged on the client side if lost from a stream).
Hence, while the apparent failure rates at a couple of points in the route is to CCP is by no means desirable, from a comms & application point of view it does not have to be disastrous, as long as the code assumes that these kinds of errors will occur and can resend or rerequest the data that was lost. This sounds easier than it is - particularly if streaming encryption is involved.
Interestingly the problem seems connected with a virtual function call (which error often displays if you close the application before the socket times you out) - "Attempted call of a pure virtual function" (or something similar) is displayed. Since the error seems to occur at one of a few points: - On login after the authenticating bar displays - On entering the game as a character - On selection of an object to approach/warp towards - On exit or entry of a station - On selection of character details, etc
In other words when there is network comms involving transmission of data to the server requiring some confirmation action by the server and result to be returned - it looks like a commonly used comms routine is at fault, and that there is an attended 'state change' in the client - exactly the scenario where the the client can not resend, or doesn't know to resend the information.
The particular problem showing on my packet analysers is not that the packet is actually lost, but that the packet has been received without all of its data - ie. segment loss. This may be confusing the application.
Obviously if the comms line has a high failure level then the problem that might not show up in an intranet shows up in production.
A couple of the suggested solutions earlier - like leaving the ping plotter running in order to 'keep the connection alive' are disastrous solutions for all users as that solution just massively increases the traffic the faulty router is handing and may single handedly cause the problem for all users as the router will rapidly get overloaded with traffic. It is quite possible that the traffic through the 26000 port is so high that it is being load balance by telecity, and load balanced incorrectly so that return routes are being lost for individual packets.
|
Carreira
|
Posted - 2009.03.24 14:06:00 -
[706]
Update:
I just got disconnected with a "socket closed" message, PingPlotter shows barely any packet loss.
|
Nyota Sol
|
Posted - 2009.03.24 15:05:00 -
[707]
I've bad bad socket problems with eve several times, but only once i left Verge to do missions in Sinq. I crashed upon entering the region and could not reconnect to eve (socket error).
The only thing that saves me after i get a socket error is rebooting my computer.
In 2 cases, i reset my router and got back into eve (although all other internet applications worked FINE). But eventually eve crashed with more serious errors and i had to reboot windows.
Those who think this is merely some remote networking or drive problem need to step back and re-evaluate. There's something very wrong here.
Between this and new crashes i'm getting, I may have to take a break from Eve.
|
Silme
|
Posted - 2009.03.24 19:53:00 -
[708]
One more person with this problem. Started for me with the Apoc patch, never had it before then.
|
Muskaos
|
Posted - 2009.03.24 20:19:00 -
[709]
Has anyone ran logserver when this happens, and pored through the results?
I cannot stay connected for longer than 15 minutes, usual mode of disconnect is chat, loading market, etc. stops responding, and 30 seconds later, client gives up and I get comms lost message at login screen.
I can't go out and do anything with my corpmates, who are all out having fun exploring wormholes...
|
john roe
Children of Gjallarhorn
|
Posted - 2009.03.24 21:44:00 -
[710]
Originally by: Muskaos Has anyone ran logserver when this happens, and pored through the results?
I cannot stay connected for longer than 15 minutes, usual mode of disconnect is chat, loading market, etc. stops responding, and 30 seconds later, client gives up and I get comms lost message at login screen.
I can't go out and do anything with my corpmates, who are all out having fun exploring wormholes...
"you are not alone" / Animatrix
|
|
Arthor Dark
|
Posted - 2009.03.24 23:35:00 -
[711]
This has been happening for a while now for me. Just today, I was checking market orders on my alt (I run 2 clients on this computer) and I see my main toon go offline (I have my main in my address book.) The alt didn't lose connection, but my main just disconnected for some reason. It doesn't immediatly disconnect either, it waits for me to do something like bring up my character sheet, which turns up blank, and then it disconnects a minute later.
Fix please, thx.
|
Hatsim
The Collective Against ALL Authorities
|
Posted - 2009.03.24 23:57:00 -
[712]
heh, i think i have had the most problems with this ****, aprox 60 times per day i get CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...
"imagine a dozen hornets pouring from the devil's mouth, Now imagine they have autocannons."
|
Grall Sek
Caldari CREO Technologies
|
Posted - 2009.03.25 00:16:00 -
[713]
:) not sure about that, im trying to leave a system for more than 2 hours with no success |
Kreel Reaver
|
Posted - 2009.03.25 01:03:00 -
[714]
About to cancel all of my accounts if this isn't fixed soon. Completely unplayable and seems to be getting worse. No mention of anything in patch notes for tomorrow either.
|
john roe
Children of Gjallarhorn
|
Posted - 2009.03.25 11:13:00 -
[715]
Originally by: Hatsim heh, i think i have had the most problems with this ****, aprox 60 times per day i get CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...CONNECTION TO SOCKET CLOSED...
QFTFT
i cant even log in on one of my chars stuck in 0.0!
all hail CCCP! (irony)
|
clapdoctor
Amarr 24th Imperial Crusade
|
Posted - 2009.03.25 13:14:00 -
[716]
Ive been having this problem during "peak" times (network drops, cant log in etc.).
Try upping your default TTL on TCP/IP from 128 to 255 by adding/updating the registry key HKLM\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters\DefaultTTL with value REG_DWORD(255). (requires reboot).
Usual disclaimer about messing with reg.
Further info - http://support.microsoft.com/kb/314053/en-us
|
Lord TYMAN
You're Doing It Wrong
|
Posted - 2009.03.25 14:08:00 -
[717]
ffs.
From the time it took me to start and finish writing a corp mail, I got DCed due to this Sock error. Lost the entire mail which was both witty and informative. I'm not writing it out again. These things are best written off the cuff, 2nd time round it just seems laboured.
So yeh, feeex eeet.
|
Grall Sek
Caldari CREO Technologies
|
Posted - 2009.03.25 16:39:00 -
[718]
Originally by: clapdoctor Edited by: clapdoctor on 25/03/2009 14:25:41 Edited by: clapdoctor on 25/03/2009 14:06:48 Ive been having this problem during "peak" times (network drops, cant log in etc.).
Try upping your default TTL on TCP/IP from 128 to 255 by adding/updating the registry key HKLM\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters\DefaultTTL with value REG_DWORD(255) and maybe try experimenting with the "KeepAliveInterval" and "TCPMaxDataRetransmissions" keys
im trying keepalive=500 and maxretries=10 at mo.
(requires reboot).
Usual disclaimer about messing with reg.
Further info - http://support.microsoft.com/kb/314053/en-us
Sorry i know very little about computers or those measures, are those settings in hexadecimal or decimal? |
clapdoctor
Amarr 24th Imperial Crusade
|
Posted - 2009.03.25 18:48:00 -
[719]
Originally by: Grall Sek
Sorry i know very little about computers or those measures, are those settings in hexadecimal or decimal?
All the values are decimal values.
For M$ XP, backup your registry (with regedit or create a system restore point), then Start->Run "regedit" and navigate to
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Properties
then, if the values do not already exist (which they probably wont), right click on the "Properties" key in the left hand pane and select "New->DWORD Value". Create three "values" as follows
DefaultTTL KeepAliveInterval TCPMaxDataRetransmissions
Double left click on the new values in the right hand pane to modify them as follows... DefaultTTL=255 (hex value 0x000000ff) KeepAliveInterval=500 (hex value 0x000001f4) TCPMaxDataRetransmissions=10 (hex value 0x0000000a)
DefaultTTL is probably spurious, but it wont hurt.
It's probably not a good idea to drop KeepAliveInterval below 500 (microseconds) as this might be seen as "storm" DOS traffic at CCP or points inbetween .
Hope this helps.
|
Grall Sek
Caldari CREO Technologies
|
Posted - 2009.03.25 20:58:00 -
[720]
Thank you, going to try it when i get home |
|
|
|
|
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 30 40 50 .. 52 :: one page |
First page | Previous page | Next page | Last page |