investigating AGM-84H has problems with the use of high-altitude flying objects in multiplayer mode with desynchronization

[RokaOvO]

When AGM-84H(SLAM) enters the dive phase, the server and the client that launched the missile are all right, and the other clients will see that the AGM-84H does not react, even if the missile launched by the client that launched the missile hits, the missile in the eyes of the other clients will continue to fly.

The missile in our perspective is at an altitude of 24000 feet at the approach stage

The distance is about 50 nautical miles

If necessary, I will provide screenshots or videos.

（I don't know English and can only use translations, maybe some of the sentences are wrong.）

[Lord Vader]

Hi @RokaOvO, your English is fine.

We would require a short track replay of the event you're describing as it's a bit hard to replicate with just your description.

Please supply a track replay of the client seeing the missile, as well as the server track, if possible. 

Just be warned that weapon desync can happen in servers running heavy scripting and/or mods. If that's the case, we ask that these tracks are recorded in standard servers running default files only.

[RokaOvO]

27 minutes ago, Lord Vader said:

你好@RokaOvO，你的英语很好。

我们需要对您所描述的事件进行简短的回放，因为仅凭您的描述很难复制。

如果可能，请提供客户端看到导弹的轨迹回放以及服务器轨迹。

请注意，武器不同步可能会发生在运行大量脚本和/或模组的服务器中。如果是这种情况，我们要求这些曲目仅记录在运行默认文件的标准服务器中。

Can I use TACVIEW?

[Lord Vader]

Sorry, Tacview is a very robust tool to analyse flight and other characteristics but we cannot use it to debug issues in DCS.

[RokaOvO]

Just now, Lord Vader said:

对不起，Tacview是一个非常强大的工具，用于分析飞行和其他特性，但我们不能用它来调试DCS中的问题。

I have over a hundred triggers on my server quest, with mods installed but nothing to do with US-made weapons, I saw what you said earlier about running a standard server quest, and I'm going to retest this with my friends tonight.

He's in class, so please wait.

[RokaOvO]

53 minutes ago, Lord Vader said:

Sorry, Tacview is a very robust tool to analyse flight and other characteristics but we cannot use it to debug issues in DCS.

 

[Lord Vader]

@RokaOvO

A track file is a recording, like a black-box file than can be saved for further analysis. When in multiplayer, it's automatically stored in "C:\Users\USERNAME\Saved Games\DCS\Tracks\Multiplayer". That's the file we need, preferably the client file and the server file of the same session. 

But again, if you have multiple scripts or mods, please try to recreate the issue without them or these track files will not replay correctly. 

[RokaOvO]

1 minute ago, Lord Vader said:

@RokaOvO

轨道文件是一种记录，就像黑匣子文件一样，可以保存以供进一步分析。在多人游戏中，它会自动存储在“C：\Users\USERNAME\Saved Games\DCS\Tracks\Multiplayer”中。这就是我们需要的文件，最好是同一会话的客户端文件和服务器文件。

但同样，如果您有多个脚本或模组，请尝试在没有它们的情况下重新创建问题，否则这些曲目文件将无法正确重播。

Hi

Before you send this reply, I understand what you mean and call a friend to start testing, and the screenshot just now is one of the bugs that have appeared so far.

In this test, I recreated the mission, as well as tested the manual entry of coordinates as well as the entry of coordinates by a third-party program, firing the weapon at the height where the issue occurred with a high throw and a flat flight throw.

Problems include altitude difference after launch, 90 degrees to the left or right after launch, flying over the target's head in the final phase, and AGM84H still flying after the explosion

For all the information, please refer to this video playback (a bit time-consuming, my friend has something to do temporarily)

Since my memory of AGM84H's control is still in an earlier version, I didn't put the weapon on the DataLink 13 for the first sortie.

server-20240607-155802.trk 测试(Player1)-20240607-160030.trk 测试(Player2)-20240607-155936.trk

[RokaOvO]

1 hour ago, Lord Vader said:

@RokaOvO

A track file is a recording, like a black-box file than can be saved for further analysis. When in multiplayer, it's automatically stored in "C:\Users\USERNAME\Saved Games\DCS\Tracks\Multiplayer". That's the file we need, preferably the client file and the server file of the same session. 

But again, if you have multiple scripts or mods, please try to recreate the issue without them or these track files will not replay correctly. 

My guess is that if the server is not able to synchronize the player's waypoint (or if it is not functioning properly)

, the host will launch the missiles according to the set situation, and the server will fix the missiles in the direction where they were just shot.

And the AGM-84H is capable of firing when the offset target does not exceed 90 degrees, and if I take off the second time (manually entered the coordinates) and do not launch the missile at the waypoint, then it will deviate as before, and then inexplicably turn around and plunge into the sea.

[Lord Vader]

We will have a look, it does indeed look strange in your tracks, but we need to reproduce it to make sure it's not just a server sync problem.

[okopanja]

1 minute ago, Lord Vader said:

We will have a look, it does indeed look strange in your tracks, but we need to reproduce it to make sure it's not just a server sync problem.

Maybe this helps, but sadly no trk/tacview from me.

About 2-3 months ago, there was a player flying the hornet. He would take of steeply with high AoA, G, climb, launch the missile and than land to refuel/rearm.

Meanwhile his missile flew in: his client, server, all other clients. The missile in his universe finally arrived and hit me (he targeted spawning slot). The later examination of the server tacview showed that Tunguska's did destroy in server universe all missiles. This was true for the I recorded as well TRK as well. So in his own universe the missile kept flying and hit me.

I do think that this desync can be triggered intentionally, and it seems that hornet client is especially prone to this (aggressive paddle use?).

Sadly I can not contribute with TRK/tacview since I deleted it, and still it's his TRK that would probably matter in this case.

About the TRKs: I observed that they do not recording 100% accurately what was happening. E.g. try driving through the grass while taxing and most likely you will never be able to get the accurate reproduction of original flight. I am pretty sure that within simulation you have some randomness in some parameters that do not get recorded within the TRK files.

[RokaOvO]

34 minutes ago, Lord Vader said:

我们会看一看，它在你的轨道上确实看起来很奇怪，但我们需要重现它以确保它不仅仅是一个服务器同步问题。

I re-called my friend to test it, and it turned out that my conjecture was correct.

In the new test, we used whether to enter exact coordinates, to fire at the target, and to launch at an off-angle angle.

Looking at these missiles on the server side, they did not receive a detailed coordinate point information after launching, resulting in these missiles will fly at the initial course and altitude, and when controlling the screen and designating a new target to the missile in the final stage, the missiles on the server side will also receive the target point information, and then the missiles will turn to the target point (those missiles that deviate will also do this) and then after the missile hits the target, the missile on the server side will lose its coordinate attitude and turn to direct flight.

The designation on the screen is to specify a coordinate value, not a unit, and we specified the missile on the ground in the final test.

For more information, refer to the new playback file

测试-20240607-180916.trk server-20240607-180909.trk 测试-20240607-180945.trk

[Lord Vader]

No more tracks required. We'll be testing this internally to verify if this is a condition on your own server or a general problem. 

Thanks.