I’ve been having constant socket closed errors when leaving grinds after destroying citadels, and also when changing grids in general it would happen randomly.
I just tested a fix by using my ISP provided router in modem mode, and that fixed the issue. So it;s my router in this case. I’ve ordered one to use instead of the ISP one (virgin media SuperHub 2ac, which are pretty bad)
Hope this helps some of you guys
Follow up question, when does the server decide to disconnect a client? Last time I tested it took over 4 minutes
Can confirm. I personally can stay logged in 23/7 if I am just in space doing absolutely nothing, or docked up in a quiet system. However, I ALWAYS experience socket closed when doing PVE content that has any kind of waves/spawns. Anomalys/DED sites or the Drifter Event site (The one you need to probe).
I use 3x Gilas from 3x Accounts to run these sites usually, and I have always without failure had at least 1 socket closed from these accounts during a DED site (it only ever happens after killing the final NPC in the site, especially if it is the NPC which completes the site and removes it from the cosmic signature list, or rarely the ‘trigger’ NPC that causes the next wave to spawn).
I can be streaming at the same time as running these 3 clients, and there be absolutely no problem with the stream 4k/s upload or better constantly no dropped frames, so it’s absolutely not my connection, and often the other 2 clients continue behaving as normal, but at least 1 client will be forced DC after completing a cosmic signature combat site. I speculate that it may have something to do with Escalations/Removing the signature
Just ran another drifter site with loglite running. I always without failure will get socket closed at the end of this site when the site disappears from the cosmic signature list. I ran it this time with 3 accounts, warping 2 characters out when the named drifters were left in the site, but leaving 1 inside the site to kill remaining event NPCs until the site was ‘cleared’ so it would disappear from the cosmic signature list so another would respawn shortly after. The other 2 clients have no issues at all as long as they warp out before the site is ‘cleared’
Here is the specific section of the log where I got session closed.
Does anyone from CCP actually read this section of the forums?
CCP is letting us down like the old days.
No answer at all and they keep expecting the money.
Not from me, not until this is fixed.
[12:22:49] EVE System > Connection to chat server lost
[12:22:51] EVE System > Reconnected to chat server
[12:26:36] EVE System > Connection to chat server lost
[12:26:38] EVE System > Reconnected to chat server
[12:27:49] EVE System > Connection to chat server lost
[12:27:51] EVE System > Reconnected to chat server
http://portquiz.net:5222/
You have reached this page on port 5222.
Your network allows you to use this port.
By the way we do not appear in local chat when we go into K-space, we could use it to get some juicy kills, be advised.
Sigh…
Tracing route to d1hoqe10mv32pv.cloudfront.net [54.230.244.122] over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms xxx.xxx.xxx.xxx
2 2 ms 3 ms 3 ms lo0.bng1.mel4.on.ii.net [150.101.32.44]
3 4 ms 3 ms 3 ms ae4.cr1.mel4.on.ii.net [150.101.33.106]
4 3 ms 3 ms 3 ms be19.cr2.mel11.on.ii.net [150.101.35.206]
5 3 ms 3 ms 3 ms nme-apt-bur-crt2-te-gi-0-7-0-13.tpgi.com.au [203.219.106.81]
6 3 ms 3 ms 3 ms 203-219-155-3.tpgi.com.au [203.219.155.3]
7 4 ms 3 ms 3 ms 202.7.162.210
8 * * * Request timed out.
9 * * * Request timed out.
10 14 ms 13 ms 13 ms 54.240.204.215
11 * * * Request timed out.
12 * * * Request timed out.
13 * * * Request timed out.
14 14 ms 14 ms 13 ms server-54-230-244-122.mel50.r.cloudfront.net [54.230.244.122]
Trace complete.
Tracing route to d17ueqc3zm9j8o.cloudfront.net [54.230.244.20]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms xxx.xxx.xxx.xxx
2 2 ms 3 ms 3 ms lo0.bng1.mel4.on.ii.net [150.101.32.44]
3 3 ms 3 ms 3 ms ae4.cr1.mel4.on.ii.net [150.101.33.106]
4 4 ms 3 ms 3 ms be19.cr2.mel11.on.ii.net [150.101.35.206]
5 4 ms 3 ms 3 ms nme-apt-bur-crt2-te-gi-0-7-0-13.tpgi.com.au [203.219.106.81]
6 3 ms 3 ms 3 ms 203-219-155-3.tpgi.com.au [203.219.155.3]
7 3 ms 4 ms 3 ms 202.7.162.210
8 * * * Request timed out.
9 * * * Request timed out.
10 * * * Request timed out.
11 * * * Request timed out.
12 * * * Request timed out.
13 * * * Request timed out.
14 13 ms 13 ms 13 ms server-54-230-244-20.mel50.r.cloudfront.net [54.230.244.20]
Trace complete.
Tracing route to dm794883twbxj.cloudfront.net [54.230.244.105]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms xxx.xxx.xxx.xxx
2 3 ms 3 ms 3 ms lo0.bng1.mel4.on.ii.net [150.101.32.44]
3 3 ms 3 ms 3 ms ae4.cr1.mel4.on.ii.net [150.101.33.106]
4 3 ms 3 ms 3 ms be19.cr2.mel11.on.ii.net [150.101.35.206]
5 3 ms 3 ms 3 ms nme-apt-bur-crt2-te-gi-0-7-0-13.tpgi.com.au [203.219.106.81]
6 4 ms 3 ms 3 ms 203-219-155-67.tpgi.com.au [203.219.155.67]
7 37 ms 4 ms 4 ms 202.7.162.210
8 * * * Request timed out.
9 * * * Request timed out.
10 13 ms 13 ms 13 ms 54.240.204.215
11 * * * Request timed out.
12 * * * Request timed out.
13 * * * Request timed out.
14 13 ms 13 ms 13 ms server-54-230-244-105.mel50.r.cloudfront.net [54.230.244.105]
Trace complete.
I don’t really see what cloudfront has to do with actual connections to the game server, though. Cloudfront is just a distribution network is it not? So web content and static content? Why would this impact connection to the live server delivering dynamic content? Why would CCP build it this way? Even CCP’s not-at-all-helpful socket closure wiki page points out that live server content is not cacheable (which makes sense).
What we really need to be doing is traceroute on the Tranquility IP. BUT, it appears that the EVE server doesn’t respond to ping queries, so we can’t reliably do so:
Tracing route to srv200-g.ccp.cc [87.237.38.200]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms xxx.xxx.xxx.xxx
2 3 ms 3 ms 2 ms lo0.bng1.mel4.on.ii.net [150.101.32.44]
3 3 ms 3 ms 3 ms ae4.cr1.mel4.on.ii.net [150.101.33.106]
4 15 ms 14 ms 14 ms be20.cr3.syd7.on.ii.net [150.101.33.28]
5 15 ms 14 ms 14 ms be-50.cr2.syd7.on.ii.net [150.101.37.12]
6 15 ms 14 ms 14 ms syd-gls-har-wgw1-be-40.tpgi.com.au [203.219.107.253]
7 17 ms 15 ms 15 ms 203-221-3-3.tpgi.com.au [203.221.3.3]
8 174 ms 161 ms 161 ms las-b24-link.telia.net [213.248.95.232]
9 264 ms 264 ms 264 ms nyk-bb4-link.telia.net [62.115.116.97]
10 309 ms 308 ms 308 ms ldn-bb4-link.telia.net [62.115.136.184]
11 306 ms 306 ms 306 ms ldn-b1-link.telia.net [62.115.143.27]
12 310 ms 309 ms 309 ms simafelagio-ic-325056-ldn-b1.c.telia.net [62.115.155.45]
13 349 ms 349 ms 349 ms intgw.danice.ldn2.verne.sip.is [178.19.49.190]
14 373 ms 373 ms 373 ms gw-int.verne.sip.is [178.19.49.217]
15 358 ms 966 ms 1233 ms 178.19.53.129
16 377 ms 376 ms 376 ms 178.19.53.130
17 * * * Request timed out.
18 * * * Request timed out.
19 * * * Request timed out.
20 * * * Request timed out.
21 * * * Request timed out.
22 * * * Request timed out.
So I’m making it to Iceland no problem (I thought Tranquillity was served from London, though?!?), at least at the time of the tracert. And I assume the time outs from the hop after 178.19.53.130 (some Icelandic ISP or data centre?) is because the Tranquility server doesn’t respond to ping.
In short… This really is something that CCP needs to work out at their end, even if it’s not a problem specifically on their network/hardware. They have the tools to problem solve this (or at least identify where the problem is occuring along the route and bust the owners balls), while us end-user schmucks do not.
PS: All of these routes are via the Google DNS. I doubt the routes would be any different if I used my ISP’s DNS server.
By the way we do not appear in local chat when we go into K-space, we could use it to get some juicy kills, be advised.
Not really. The game server knows you’re in local, as does anyone that is connected to the chat server. It’s just you who can’t see yourself (or anyone else) in local due to loss of chat channel connection. So it could lead to some juicy losses, as opposed to juicy kills.
This pretty much sums up my awesome gameplay experience. 2 clients connected, 1 drops, yet the other one stays connected no problem.
Hello here,
I can’t seem to log in to the game with my second account… I’ve been having multiple socket closed issues from time to time past few months, but today it just wont start at all.
I logged in with my main account no prob, but the alt account just wont start.
I have the socket closed issue when launching the game from launcher.
Sometimes it gets through the character selection, but when I select my character, long loading screen, and socket closed.
What’s weird is that the character appears logged in ingame from my main (I have the pop up), but when the socket closes on the other client, the character stays logged in for tywo minutes before being set offline.
I just cant play with my second account this morning, what should I do ?
Returning back to EVE from a few months pause and for the first time since I started playing this game back in 2005 finally I got to experience this “socket was closed” while idling in stations and random disconnects/connects to chat windows.
Last night I got tired of this issue and sent in a ticket with all kinds of info, dxdiag, network diagnostics, ISP, internet speed tests and today I ofc got this copypasta answer everyone gets!
Living in western parts of sweden, ISP is ComHem 250/50 maxing out those speeds and a friend of mine in the same city, ISP Telia having the same issue… Well ofc it was probably my pc, hardware or ISP that was the problem. The fact that other friends with those ISP’s in or close to our city dont have any issues, well… and its only EVE having this problem, no other online gaming!
After many hours testing, tweaking, troubleshooting all kinds of stuff I finally made progress as it looks! Now my character has been online almost 4h and counting instead of 3-4 minutes! No more disconnect from chat windows.
Solution, VPN and connect to Norway - Oslo, Akershus county 
I think someone suggested it in another socket closure thread that maybe CCP is running some kind of automated anti-bot tool that is a little (a lot) over-zealous in however it does it’s magic, and is kicking off legitimate players (usually running multiple accounts).
It wouldn’t surprise me if this was the case.
im having same problem 3 days in a row socket closed at 12 lunchtime today closed at 11am. 1st time it happened tried to get back in every 5mins failed for hours. then the rest of times waited for an hr then logged back in no problem n played for hrs with no problems. nearly time to see if that stratergy works today.
Socket closed on me 17 times in about 30 minutes. Just gave up and left. Eve Online really is unplayable.
Any updates on this issue? I have filed a ticket but not much going on there. I am having the same issue, one client will stay on no matter what, while a main gets closed if I look away for 5 minutes, sometimes while I am sitting at the computer interacting with the Eve client.
I still get mass socket closed when reinforcing citadels with multiple accounts. They are fine and only ever disconnect at the point when the citadel dies or changes to a new reinforcement state, the exact same as listed above for completing NPC sites.
Same here. Problems started today at 10:30 Eve Time. Was unable to log in via normal wi-fi with socket closed errors instantly after the game launched.
Phone hotspot worked though for some reason.
I switched isps and added a vpn. Problem solved.
I’m having this issue too, and not with multiple clients running side by side, etc. I only have one account, and therefore only ever 1 client running, and all I’m ever doing is playing the game lol. Sometimes it’ll be fine for an hour or 2, but, for odd 30 minute chunks, it’s unplayable, like the last 30 minutes. And I have zero intention of switching ISPs for 1 game, ALL of my other games/internet usage work/s without any problems
I have now had repeated socket closures for the last hour. every few minutes. anyone else getting this?
is there a workround?
