Cannot run a session now with crashes/connection errors/maps disappearing

[Copernicus219]

So whatever was done to "enhance" the network, has pretty much made the game unplayable now. I ran a session and I crashed and then people were having errors - license key conflict, something about name servers, etc. We would get everyone in to the game except 1 person( different people each time) then I'd have to restart the server again which takes like 5mins easily. Then after about an hour of this, finally got everyone back into the game and my map was just gone, when I opened it, it was grey. So now I have to redo the map, adding all the LOS. So killed the game session.

31644

[Copernicus219]

After 5 mins of launching:

31645

[Zacchaeus]

That looks like you don't have the latest update.

[Copernicus219]

That looks like you don't have the latest update.

Not sure how you came to that conclusion, but yes, I have the latest update.

[Sgain]

same here on my LAN - never had this issue before today.

[pollux]

So whatever was done to "enhance" the network, has pretty much made the game unplayable now. I ran a session and I crashed and then people were having errors

It's interesting to report that things didn't go well, and with Sgain corroborating maybe the report will help the devs figure out what's going on, but to REALLY help the devs figure out what's going you should collect the error messages people are having and paste them into your report. I know that requires stopping the play session, but it's also how to make fixes happen quickly.


So killed the game session.

Some advice I gave in another thread:


The latest guidance on live play sessions for fun is still "don't". You risk session breaking bugs and campaign corruption: https://www.fantasygrounds.com/forum...l=1#post471093. That said, it's been viable for a while to run TEST one-shot sessions locally as long as your players know they're being guinnea pigs. And with yesterday's update, the networking has received a major update as well. We may now be in a state where networked test sessions work well enough for real playtesting.

In my opinion, the way to decide if FGU is ready for you to run a session is to ask yourself how you'd feel if you had to stop the session early due to a game-breaking bug that you couldn't play through and had to stop, characterize, and report immediately. If that session feels like a success because you found and reported an important bug, and finding bugs is why we're here in the beta... then FGU is ready for you. If you feel angry or disappointed about FGU interrupting your game... then run it in FGC which is a known quantity.

If it's possible for FGU to "kill" a session in your mind, you shouldn't be using FGU to run that session in the first place. FGU is a good match if a "successful" session looks like "We hit a bunch of bugs, we stopped playing to record the error messages and post our experiences to a forum thread. It was a really productive testing session!" If a successful session for you means not hitting bugs, you should be running the session in FGC.

[Copernicus219]

I would love to run in FGC, but my router won't let me because of something to do with flood control and port forwarding issues, which is the reason I decided to dump $135 into FGU. I did ask, but no one got a ss of the error. I realize this is beta, but this stage beta shouldn't be show-stopping bugs. I do QA for a living, and while I don't mind helping when I can, FGU is now a month and a half behind schedule. The reason I say it killed a game session is important because imo all the big, showstopper bugs should have been squashed by now. The dev's need to know that for them to meet their new goal in 1.5 months that the game is still so unstable to even play more than a half hour. I'm not being paid for this QA work. I will bring as much info I can when I run into bugs, but it's not my job to do this, it's the dev's job.

[pollux]

You sound frustrated and like FG isn't working for you right now, which sucks. I'm sorry things aren't going smoothly. It's certainly not your job to be debugging FGU if you don't want to be. If you're using FGU for live play today though, you are volunteering to make its problems your problems. It's marked as a beta and the dev team have stated in no uncertain terms that it's not ready for live play. It's reasonable to WANT it to be ready for live play. It's reasonable to note that it's already been delayed and to say that it SHOULD be ready for live play by now. But that's not the reality, and it's mostly not under our control. If you double-down on running live play sessions in FGU today... you're definitely choosing to have more frustrating/failed sessions. None of us... not you, the community, or the devs (who are already working as fast as they can) can prevent that.

... which is why I suggested FGC. It sounds like you've already tried that and already faced frustration there too... which sucks doubly. However... it is POSSIBLE for you to sort out the networking problems in FGC more or less immediately, and more or less on your own (perhaps only after investing an unreasonable amount of time and irritation and requiring an unreasonable amount of community help). Having to work so hard to get any game running sucks triply... but with FGC it's at least POSSIBLE to get a stable game going today, doing the same with FGU requires time and the cooperation of many people characterizing and fixing the bugs that unfortunately do still exist. So for live play, I still recommend you circle back on FGC and fix your networking problems, which can be overcome today (unlike your FGU problems). The next place to check out if your router has already failed all the common advice around port-forwarding is this: https://www.fantasygrounds.com/forums/showthread.php?43607-Port-Forward-Alternatives I personally would look to Hamachi next as I believe it can be free for most use-cases and it and doesn't have any router setup requirements.

[Lilkev716]

I started having that issue as well. However I found going into settings and re-logging in with credentials before starting server fixed the issue for me.

[LordEntrails]

Be aware, the entire network library for FGU was replaced in the last week or so because the previous one wasnt working. So yea, it's frustrating, but the network side of FGU is really brand new and bound to have issues.

[Moon Wizard]

@Copernicus219,

I'm not sure why your connection is dropping; I've actually had the network library developer look at it too. Neither of us have been able to recreate the dropout. I'm going to try again tomorrow to recreate from my machine.

For additional information:
* Can you PM me a copy of your traceroute from your machine to noblewhale.com? (which is the developer server for matchmaking services)
* What is the make/model of your router?
* What security software / internet filtering / parental control software are you running on your machine/network?

Thanks,
JPG

[Moon Wizard]

Also, does this seem to disconnect only when someone is connecting? If so, can you check your router for flood protection settings. Also, if so, do you happen to be using Xfinity? (which made some interesting changes recently to their routers)

Thanks,
JPG

[Copernicus219]

Nope, using AT&T fiber.
just launched and after a few minutes got the error:
31672

this was with no one connecting.

[Copernicus219]

Pace - 5268AC

[Trenloe]

I would love to run in FGC, but my router won't let me because of something to do with flood control and port forwarding issues...
Can you expand on these issues? These may effect FGU as well.

[stewartl42]

I, too, have these issues on AT&T Fiber, and I have the same router. I used to work for AT&T. They pushed out a firmware update that enabled a bunch of DDOS countermeasures, one of which is "TCP Flood" detection. If it detects too many connections from the same IP address, it drops them and throttles the connection. No way to disable them.

I'm a Network Security engineer by trade, so I set up a Virtual machine in my home network and built an SSH tunnel that my players connect to, then run their game through the tunnel. Because it's encrypted, the router never sees the TCP traffic directly, and does not detect any floods. I'll admit that my solution might be overkill for many people though.

[Copernicus219]

I, too, have these issues on AT&T Fiber, and I have the same router. I used to work for AT&T. They pushed out a firmware update that enabled a bunch of DDOS countermeasures, one of which is "TCP Flood" detection. If it detects too many connections from the same IP address, it drops them and throttles the connection. No way to disable them.

I'm a Network Security engineer by trade, so I set up a Virtual machine in my home network and built an SSH tunnel that my players connect to, then run their game through the tunnel. Because it's encrypted, the router never sees the TCP traffic directly, and does not detect any floods. I'll admit that my solution might be overkill for many people though.

I had these issues on FGC and was hoping FGU wouldn't run into this because there is no port forwarding, but sounds like that may be the case? The fiber is so good though. I don't understand why I can't just use my own router/modem with ATT. I'm guessing there is something special with a fiber connection? Not sure what I can do, but I'm guessing I won't be the only one with this issue for anyone forced to use shitty routers...

[stewartl42]

I had these issues on FGC and was hoping FGU wouldn't run into this because there is no port forwarding, but sounds like that may be the case? The fiber is so good though. I don't understand why I can't just use my own router/modem with ATT. I'm guessing there is something special with a fiber connection? Not sure what I can do, but I'm guessing I won't be the only one with this issue for anyone forced to use shitty routers...


Were you running a Cloud game?

[Copernicus219]

Were you running a Cloud game?

Yes

[pollux]

I, too, have these issues on AT&T Fiber, and I have the same router. I used to work for AT&T. They pushed out a firmware update that enabled a bunch of DDOS countermeasures, one of which is "TCP Flood" detection. If it detects too many connections from the same IP address, it drops them and throttles the connection. No way to disable them.

For folks who are not savvy to SSH tunnelling, do you have an opinion about which of the commonly recommended alternatives are likely to behave well with flood detection? https://www.fantasygrounds.com/forums/showthread.php?43607-Port-Forward-Alternatives

I'd normally recommend Hamachi, but being P2P perhaps that would still be prone to triggering flood detection. PureVPN seems like it would almost certainly work well but not be free.

[stewartl42]

ZeroTier looks simple and free. And it's open source, if that means anything to you.

My problem with these, is they require installing a VPN client. At least one of my players is using a work-assigned laptop, and corporate environments frown upon using unapproved VPN clients, as a general rule. The SSH Tunnel does require PuTTY, but it's not as deeply integrated into the system as a VPN Client would be (adding virtual network cards, Etc).

The SSH Tunnel option on the list you linked to is pretty similar, but I'm running the Virtual machine internally on my network suing vmware, as opposed to in Amazon's cloud.

[LordEntrails]

PureVPN is what I use (with dedicated IP option), and it does not require the players to install anything. Totally transparent to them.

[stewartl42]

That's a paid option, right?

[LordEntrails]

That's a paid option, right?
Yes

[Moon Wizard]

@Copernicus219,

I'm reaching out to the developers who built the networking system to see if they have some ideas. Basically, it seems like something in between your machine and the network matchmaking servers is blocking or killing one of the connections required for the networking system to work.

I'm not seeing any other reports like this right now (with an admittedly small sample size), so my first guess is that it's something related to the network setup or router that is causing the issue. I'm just not sure if it's something that can be fixed by a setting or whether it requires something else.

I'll be back with more questions as I investigate. Please let me know if you figure out anything on your side.

Regards,
JPG

[stewartl42]

@Copernicus219,

I'm reaching out to the developers who built the networking system to see if they have some ideas. Basically, it seems like something in between your machine and the network matchmaking servers is blocking or killing one of the connections required for the networking system to work.

I'm not seeing any other reports like this right now (with an admittedly small sample size), so my first guess is that it's something related to the network setup or router that is causing the issue. I'm just not sure if it's something that can be fixed by a setting or whether it requires something else.

I'll be back with more questions as I investigate. Please let me know if you figure out anything on your side.

Regards,
JPG

This is easily reproduced, and I'm happy to help diagnose.


Also, Apparently some users are having luck asking AT&T to swap their router for an Arris BGW210-700.

[Moon Wizard]

Thanks. I actually have AT&T too; and I was able to recreate the issue after about an hour of letting FGU sit open at my place. So, hopefully, I'll be able to get some feedback from the network developer on how to pinpoint the root cause better.

Regards,
JPG

[Copernicus219]

Thanks. I actually have AT&T too; and I was able to recreate the issue after about an hour of letting FGU sit open at my place. So, hopefully, I'll be able to get some feedback from the network developer on how to pinpoint the root cause better.

Regards,
JPG

I really appreciate you taking the time to look into the issue! I'm sure the amount of people that are going to run into this problem is small, especially when it's a GM only issue. If there is a way to correct, it would prob work for FG classic as well. Thanks!!

Edit: I'm trying to run a session every week so I'll let you know if anything changes

[stewartl42]

Were you running a Cloud game?

Yes

I'm like you, I thought a cloud game would change that behavior. As I remember, it was the volume of GET requests to fetch game content, but I suppose that is still hosted on the GM's machine, not on the cloud Broker? Maybe a dev can answer.

[Moon Wizard]

That doesn’t appear to be the issue here; but I’ll keep that in mind. The matchmaker connection is dropping while the game is apparently idle. So, I need to get more information on what is happening internally. I’m hoping the network library developer has some ideas to dig in.

Thanks,
JPG

[stewartl42]

That doesn’t appear to be the issue here; but I’ll keep that in mind.

Maybe my issue is not the same as the one you are seeing then. and Granted, mine was on FGC, not FGU.

[Copernicus219]

So I guess I'm wondering why would flood detection have to do with disconnects when there is no one connected. I opened FGU a few minutes ago (waited the 3min+ time for it to load) and clicked on images&maps and it the disconnect popped up. It just seems really random to blame flood detection as well, because I did host a session about a month ago in FGU where we didn't disconnect until like at the end of a four hour session.
31715

[Moon Wizard]

I was able to recreate locally on my laptop, and I'm working with the network library developer to figure out the situation. It doesn't happen on my desktop, even though they are on the same wireless network.

Regards,
JPG

[Copernicus219]

Hmm, do you think it's a network card issue? My card is:

Edit: It's a killer 1535

[Moon Wizard]

I wouldn't think so, but you never know what the issue may end up being. I just looked up the adapter on both my laptop and desktop; but they are both Killer Wireless. (1535 / AX1650x) Only my laptop has the issue, as far as I know. It's harder to create here for me, because it takes an hour to trigger. That's why I'm working with network library developer to get help on tracking down.

Regards,
JPG

[Moon Wizard]

@Copernicus219,

The build I pushed yesterday included a new version of the network library with an updated timeout scheme for the relay connection issue you were seeing. On my laptop that always disconnected around the 45-60min mark, I ran an 8-hour and 2-hour session without disconnect. Can you let me know if you're still getting disconnects in the latest version, and any console messages if you do?

Thanks,
JPG

[Copernicus219]

I was on for around an hour last night by myself with no disconnect. Going to try to run another game this weekend.

[Copernicus219]

Just wanted to report, GM'd a session for over 4 hours with no one disconnecting, so whatever you guys did, thank you!!

[Moon Wizard]

Thanks for reporting back; I appreciate it. You were the main one having the issue (other than my laptop), but I'm sure more people would have been affected as we opened up release.

Regards,
JPG