Rare certificates issues on EOS 1.17.1.3

Hi, a small portion of our playerbase on Windows are experiencing a Websocket certificate issue.

The logs read:

```

Opening Lobby Socket To: “wss://api.epicgames.dev/lobby/v1/54fd5f945e76457c973a8addaa0de98b/lobbies/connect”

Lws(Error): SSL error: unable to get local issuer certificate (preverify_ok=0;err=20;depth=2)

```

We are on EOSSDK Version 1.17.1.3-44532354, so the root certificate issue should be OK, and players are actually able to login to EOS.

I’m wondering whether this issue was known on your side and if you had any recommended action?

Thanks,

Hey, that’s the error we would expect to see from players without the AWS certificate, and running a game that is using an older EOS SDK. Is it possible players still have an old version of the game on an older SDK? Could we get a full set of logs from these players? You can also ask them to run the curl command below in command prompt. That will force pull the certificate and should unblock them (1.17.1.3 should be doing this for them).

curl https://account-public-service-prod03.ol.epicgames.com/account/

Thanks!

I confirm it happens on 1.17.1.3-44532354

runing the curl command updates ther certificate (note that it returns a 404)

They get a 404 on the curl command? If the certificate wasn’t properly pulled, they are likely still seeing issues in your game. Do you know if this is affecting players of a specific region? If they can’t hit our https://account-public-service-prod03.ol.epicgames.com/account/ endpoint that points to a DNS or firewall issue…

OK some users reports that the curl command fails with CRYPT_E_NO_REVOCATION_CHECK

I told them to retry with --ssl-no-revoke

It seems these users are on a restrictive network. Could they share more info on where they are trying to play the game? They’ll need to ensure their firewall isn’t blocking access to the endpoints listed in the KB below.

Thanks!

I’ll let you know how things evolve.

I guess you know how difficult it is to get meaningful information from users:)

Hi! I’m in nbrigand team (his currently off so I follow up the situation), we managed to dig the issue with some players, here is what we tried / found:

  • No issue running the curl commands even hitting the same endpoint as the one the game hit and fail (https://api.epicgames.dev/)
  • It has also been confirmed to be happening on a latest windows 11 (we tried with the exact same windows version and it worked fine, so I think we can rule that out).
  • The players with who de dig it are on a normal house internet, we also had one of them try with a hotspot on his phone to see if it was linked to his ISP, but he still had the same issue
  • We had them disable their firewalls and antivirus, but the issue persist
  • One of the players reported us that he also has connectivity issue (cannot connect to epic servers) with overcooked 2 (which I believe also use EOS), but none on rocket league (maybe those two games do not use the same endpoint?)
  • Even tried to manually install the certificate on the machine, to see if that would help but it did not change anything.

Does that give you some direction about what could cause this issue? And if you think of anything else that we may try to debug this please feel free to tell us.

Thanks!

Victor

Hey, Are these players using a VPN? If they are not, do you know which country they are located in?

Hello Seb,

They tell us they have no VPN and I have the location of two of them: Portugal and France

I’m honestly stumped. Do either of these players have the Epic Games Store installed? If they do, are they experiencing any issues with the Epic Games Store Launcher? If they are experiencing issues with the launcher, there may be other clues in the launcher logs. It would also confirm that the issue is accessing some endpoint on our end, rather than an SDK issue.

I’m double-checking to see what could be different with Rocket League.

Thanks, I’ve asked them for the epic launcher and will let you know as soon as we have the info. Yeah on our side also we don’t really understand what could cause that, if it was only one player I would say it’s just some strange setup he has, but the fact that it impact multiple players make me think there is a real issue somewhere, but after all the tests we have done with them we still can’t find why it is happening, or what’s the common point between all those users.

Thanks for the update. In addition to the Epic Games Launcher question, can you ask them to navigate to https://api.epicgames.dev/sdk/v1/default?platformId=WIN in their browser? They should see a very large JSON, and shouldn’t see any certificate warning. If they can share a screenshot that would be helpful. Thanks!

Hey Seb, we unfortunately have a few users who are running into a similar issue. We’ve walked one of them through the recommended fixes but they’re still running into the certificate issue. I had them visit that webpage and have attached the resulting JSON. We also asked them about the Epic Games launcher and they reported that everything worked as expected as far as they could tell.

Hey, thanks for having them try. I’m doing some research as to what else could cause this error. I did think of something you can ask them in the meantime. Can you please have them confirm their Windows system clock is showing the correct current time? If there is a large enough offset, that could explain the error.

Hey, sorry for the message spam, but I just saw a last reply from one of our players, he manage to make it work by completely wiping his PC and re-installing windows.

Hey, sorry for the delay. Our offices were closed for 2 weeks over the holidays. That is interesting info. I’ve reached out to our SDK engineering team for assistance. I’ll let you know as soon as I have more info. Thanks!

Hey, we have a script that was shared with us from Microsoft. Can you ask affected players to run the script and share the output? You can download the script here. Thanks!

Thank you! Can you ask this same player to run nslookup:

nslookup api.epicgames.dev

What I’m seeing in this trace is that our endpoint is being resolved to a private ip address. That hints to some DNS, anti-virus or firewall issue.

Thanks,

Seb

Thanks, I’m talking with our EOS SDK engineering team for some additional help.