Random Quick Zoning Disconnects

Someone in our NAGD Discord posted about crashes today. It seems like this is a new phenomenon that possibly started last reset. It isn't a classic disconnect.

The issue presents as failing to zone. If you have other group members, they just see the member drop instead of zone. The member can immediately reconnect instead of going through the classic wait for LD.

If you don't notice that your member dropped, you'll eventually see the window go back to the server select screen instead of fully crashing or throwing an error.

It turns out a few of us are experiencing the same issue.

Edit: I'm a Windows 11 user.

 

I am experiencing this issue. For me, the character hangs in the zoning process and I have to manually close out the client. I am able to log back in immediately. This has happened zoning ant least in the following ways: into timeA from tranquility, zoning out of pofire into tranquility, and gating from karnors into tranquility.

edit: this is also a new issue for me in the past week or so

 

Oh, this was happening to my wife last night during our raid.

Edit: Upon further discussion, it maybe only happened to her when zoning into Time B, using the Fire portal. She was able to zone from Fire into tranquility, from tranq into time A, but then crashed when using the Fire trial portal. Others may have been fine?

 

This is not new. This has been an issue since I came back months ago. Seems most likely to happen if I have been online a long chunk of time. It is an almost certain crash on zoning. No disconnect, just windows error, and can immediately reconnect.

 

posted in the changelog thread before I saw this, sorry

This is a new issue. I'm seeing very frequent LDs when zoning. Sometimes this happens like 2-3 times in a row when zoning. The client does not crash, it's stuck zoning. Happens everywhere for me, last nite when running to ToV I had it happen like 5 times porting to CS, succoring, zoning into SG, succoring, zoning to WW, going to ToV. It's pretty bad.

There is no client crash at all - I am running debug mode on the clients (dbg-xxxx.txt files) and there's nothing in there. No crash, nothing telling.

On other clients the char drops group and gets kicked out of the game pretty quickly. So I have to manually close the game client, relaunch it and log back in. Usually happens within ~20 seconds, so it feels like a zone packet is getting dropped and character is getting booted from the world or something like that. Basically same as if they just camped out...

I've reset my router multiple times but it doesn't seem to help.


Feels like a regression of this issue, except previously the timeout was really long and just got stuck for minutes.
https://www.takproject.net/forums/index.php?threads/3-18-2020.15870/page-2#post-120239
In this case, the char is booted in like 20 seconds on zoning and can relog back in right away.

 

The thing I am describing is new. There is no windows error. Several guildies have recently experienced it as well after months of crash-free zoning.

 

tested using a VPN to Canada , didn't have any zoning disconnects today in several hours of playing

which suggests this could be a networking issue rather than a server code issue
since it's affecting many players from all over the place, the issue could be with the datacentre host ISP

maybe try restarting the internet link / router / firewalls etc? there's patch maintenance coming up soon, could do it during that downtime


edit: disconnected from the VPN to test, got a zoning disconnect within a few minutes

 

That's interesting, your VPN may be converting UDP to TCP until it leaves the tunnel. Interesting this could be some larger provider issue with UDP or perhaps something happened in Windows in a recent update which happened last week. Just tossing out ideas.

 

This is exactly the issue from which I suffered for months, at least in the way its presenting. 50% or greater chance of getting stuck zoning, but could close the instance and relaunch immediately without being caught in LD mode. All that you guys describe is exactly what was happening to me. It was remedied entirely when I changed ISP and so lay somewhere within the network I guess.

May or may not be helpful but the symptoms are identical.

Mac OS, Mac client, no current issue

 

I've also been getting a similar "disconnect while zoning" thing on a Mac for the last few months. If I'm three-boxing and zone ten times, odds are pretty good that at least one toon will disconnect on one of the zones. Sometimes I'll get 2-3 disconnects during a raid with lots of gating (okay fine, lots of dying), other times I'll go a week with no disconnects.

Some netcode does seem to catch these particular disconnects and punt the toon from the server, so you can reconnect quickly without having to wait for the toon to go linkdead and eventually disappear. That netcode makes these disconnects a lot less annoying, and also suggests that someone (Haynar?) has worked on this problem in the past.

Anyway, not super specific information, just noting that this issue is not unique to Windows and not new since the last patch.

 

This issue is definitely new for me.

Upon further reflection, this issue did not start occurring after last patch. It started happening mid-way through the patch, around Sunday/Monday 7/8th June for me.

I was camping a named which required pulls to zoneline to snare park it. This required zoning around 10-15 times per pull (between the three characters). No issues in the first week. In the second week one day I just started timing out when zoning which made the pull insanely frustrating. I restarted my router a few times and couldn't solve it, so I stopped camping the named because the disconnects made it insufferably annoying.

The issue did not resolve itself after a few days so playing became much more frustrating. I have to use a VPN connection to play, which causes issues with my other normal computer usage (all my other apps freak out due to connections from another country, and everything is slower to browse). So yeah, having to switch between VPN on/off to be able to play now makes this much less palatable.

 

Hello,I had 6 disconnects last night zoning between various zones. The issue appears to be getting worse.

also, can support the assertion that the issue is new. I am aware of the disconnects/crash that may happen when a client has been online for a long while, and these disconnects are unrelated to that. They happen regardless of how long you’ve been online.


Zone > within a few seconds the toon will disconnect (drop group if in group) > client will need to be restarted for fastest remedy.

The account is able to log in immediately, the character is not LD stuck zoning.

 

I posted in another area no aware of this thread (thanks Rimi). I am getting disconnects as well. mostly 1 box drops out (no matter if i am boxing 1, 2, or 3) though I did have one this morning where two of my three dropped just like this. Had it happen to me about 8 or 9 times this morning (was doing chanter epic so a lot of zoning) and seemed to drop at least 50% of the time when I zoned.

 

For me it is usually the same box several times in a row. I have had it occur with two boxes at once, but that only happened one time for me as well.

 

seems awfully close to the issue I posted https://www.takproject.net/forums/index.php?threads/random-client-disconnects-no-linkdeath.38362/

You could just use split tunneling to only affect the eqgame.exe process if your vpn has a GUI or set your vpn conf file to isolate connections to 70.35.159.18 e.g. "AllowedIPs = 70.35.159.18/32", so the VPN only routes traffic going to / coming from that IP.

 

thanks for suggestion, will keep trying to make split tunnelling work so I can play, having some challenges with it

as another test, using same VPN client just choosing a different server location (eg, not Canada) results in same disconnect issues

so it doesn't sound like my ISP issue, and it's not just the property of the VPN connection encapsulation that fixes it, but rather changing my network route to specifically be closer to Canada for the client->server connection distance

so again, this feels like a TAKP host datacenter ISP issue

 

Troubleshooting results using same computer, same base internet connection, same VPN client:

native internet from NZ (no VPN) - yes disconnects
Australia VPN server - yes disconnects
Canada VPN server - no zoning disconnect issues
Utah VPN server - no zoning disconnect issues
Oregon VPN server - no zoning disconnect issues
New York VPN server - no zoning disconnect issues
Florida VPN server - yes disconnects
Texas VPN server - yes disconnects

you can immediately tell when it will disconnect because when you hit the zone, the packet that makes your UI elements disappear does not get received. the client freezes on the zone, but the UI elements disappear around 15 seconds delayed. if you leave your client running, it will eventually unfreeze back at server select screen around 1 minute later and you'll be able to click right back into char select screen because your character would have been booted form the world

here's a video of what it looks like

 

anyone who is having this issue on TAKP also play on p99 also?
are you getting disconnected there too?

https://www.project1999.com/forums/showthread.php?t=448957

 

This reminds me of a fix I believe that Haynar put in a long time ago. Randomly a 2nd or 3rd box would drop while zoning. If somehow new code is messing with that code or it wasn't included.

I will say I am relatively close to where the TAKP login server is and not having issues. If something is happening during long paths to the server.

 

Not using a VPN, this issue is still affecting my TAKP installs on both Bazzite and Win10. Last random disconnect on multiple toons while zoning was 06/26/2026 430pm AZ / PDT from PoTranq --> DSP, then DSP --> UP.

 

Just to add to the data gathering. This is happening to me every 5 or 6 zones on multiple accounts while boxing. It is as described where the window freezes and will update VERY slowly and eventually crash out. If I quit, I can reconnect immediatly. This is the latest client on Linux.

NAME="Linux Mint"
VERSION="22.3 (Zena)"

 

I am having the same issue!
I never had this issue before and I didn't make any changes on my TAKP config or folder recently.
I do not use any custom UI and the only custom mod I use is zeal.
One thing I noticed is that zoning screen sticks for longer than usual and once crash happens character is booted from game immediately instead of going LD.
This behavior makes it different than other crashes for me.

 

I think you will need to ask for help from your server host Rogean to troubleshoot this issue.

This does not seem like an eqemu server code issue. It seems like an issue at a lower layer of the hosting stack.
It could be one of these things:
1. linux operating system software (kernel, drivers, OS config settings)
2. linux virtual machine container
3. virtual network interface
4. physical host hardware
5. physical network interface (network cable, switch)
6. outbound router / DNS configuration / routing table / traffic inspection or QoS policies
7. inbound / outbound ISP

If this was an eqemu code issue then it would affect everyone on the server. It does not affect everyone.

Few questions around how the TAKP and TLP are hosted:

What is the difference between how TAKP prime and TAKP TLP are hosted and served? Which part of the stack is shared, and what is different. I think you need Rogean to help draw a network interface diagram and point out everything that might be different.

In this post, it sounds like you use the same host: https://www.takproject.net/forums/index.php?threads/server-maintenance-2025-10-13.35770
But this issue does not happen on P99, and it does not happen on TLP server.

It feels like an OS / physical layer issue, or a networking issue somewhere in the hosting stack. At least I hope. If it's your ISP then it would be more difficult to solve. But I doubt it's your ISP based on troubleshooting and the fact it doesn't affect everyone.

So questions on differences between TAKP and TLP:
1. Do TAKP and TLP use the same version of linux OS, same kernel, drivers, configuration?
2. Do they use the same VLAN interface?
3. Do they have the same outbound traffic routing / DNS / traffic shaping policies?
4. Do they share the same network cable, network switch, router?
5. Do they share the same outbound IP?
6. Do they share the same ISP?

Need to find which part of the hosting stack is different between TAKP/TLP/P99 to find the root cause

If we can't find the root cause, then you might be able to try and mitigate severity of the problem through eqemu zoning handshake code changes by making it more resilient or something (change the way the packet is sent/received, implement more forgiving retry code, etc) but I don't think this is the root cause since the issue started happening outside of server code changes.
As mentioned, it started happening for me mid-patch cycle, and the fact I can change the behavior by just slightly alternating my network route suggests it's not eqemu code issue.
It's also not a client issue since it happens for so many different people running Mac, windows 10, windows 11, different linux builds, and the severity of the issue increased around similar time for those affected people.

 

This is a very strange issue because of how specific it is (only a specific client action - zoning), and yet how widespread it is (across geographies, and client configurations).

There are no random disconnects during gameplay - I don't just start timing out while I play in the same zone, or on a raid. My trader stays online for week+ no problem.

All my clients don't drop. Only one client drops while zoning.

If this was a host ISP CDN or BGP issue, it would affect everyone. It does not affect everyone.
If it was a host ISP issue it would also affect P99 / Proj Quarm / TLP since I assume they're all hosted with same ISP?

If this was a server load or port exhaustion issue it would happen more often during 'peak times', but it doesn't matter, you get disconnected any time of the day. It would also not vary based on people's traffic route (geo locations).

If this was a eqemu code issue, it would happen right after a specific code change patch, but it started happening mid-patch for me.

If this was a client issue it would happen only for those with specific TAKP client config - but it happens for all sorts of different clients, and it didn't happen after client changes were made. If it was a client config issue I would disconnect no matter what VPN server I used, but I can change behavior by just picking a different VPN endpoint.

It is also not a transient issue, it's happening for specific people constantly no matter how many times they reboot their PC or home internet and it's been happening for weeks and seemingly spreading to more people and getting worse. It is not individual TAKP client, client PC, network router, WiFi or ISP issue.

So if Proj Quarm / TLP / TAKP all share the same underlying eqemu netcode, physical hosting location and ISP, then the issue is somewhere between the eqemu code and the outbound ISP. It's somewhere in the linux virtual or physical host layer, or in the networking config or cabling, or packet routing to specific subnets.

Something somewhere is dropping a very specific UDP zoning packet and seemingly nothing else. And it only happens for very specific traffic routes.

If this issue was happening on p99 or project quarm there would be a lot of noise about it since there's 20x more humans playing there. I assume there's no disconnect issues happening for these two servers which are apparently hosted in the same geographic location.

This issue isn't happening on TLP server for me from my limited testing. I spent an hour trying to get a single disconnect without luck. I can replicate a disconnect on TLP within a a couple of minutes of trying. Is anyone playing on TLP getting zoning disconnects?

Since I think it is likely that TAKP and TLP are most similar, next troubleshooting clue has to be: what is different between how TAKP and TLP are hosted?

 

Issue still persists after latest server reboot.
Just a side note, it happens randomly and not always.

 

Months ago when I had this issue (mac os), it was only on takp and not on TLP, just like described here. I never did find the exact issue, but I did change isp and the issue stopped for me completely. It had plagued me for many months.

In my limited experience it seems that whatever issue existed with my ISP did differentiate between TAKP prime and the TLP.

 

As others are saying - Issue is still ongoing after yesterday's reset.
Not sure if it helps, but sometimes after the crash I'm getting a pop-up error:

Exception code: 0x0000005
Exception String: EXCEPTION_ACCESS_VIOLATION
Exception occurred in module: C:\TAKPv22\eqgame.exe
Zeal Version 1.3.8 (32ad746)
Character: (Character Name)
UI Skin: UIFiles\Default\
Zone ID: ffffffff
Game state: 3
Callbacks: ExecuteCmd : Exit (0x0)
SpawnInfo: 0x0
Self 0x0

Not sure ^ this ^ has anything to do with the zoning error (I'm not technical), but I noticed it pop up yesterday so I took a screenshot of it. Normally it just crashes and then I exit and re-enter. Next time it happens I'll see if any messages pop up like this.

For reference, I'm on an iMac Pro but playing through bootcamp so native Windows 10 install