Actium

Having a look at the uploaded dcs.log, this doesn't look like a local network issue. All occurrences of ASYNCNET in the log (PowerShell: Select-String ASYNCNET dcs.log): dcs.log:21:2025-09-06 18:29:43.859 INFO ASYNCNET (10472): Login success. dcs.log:22:2025-09-06 18:29:45.017 INFO ASYNCNET (10472): Got auth data. dcs.log:26:2025-09-06 18:29:45.100 INFO ASYNCNET (Main): ProtocolVersion: 367 dcs.log:27:2025-09-06 18:29:45.100 INFO ASYNCNET (Main): Adding LAN search interface 0: 127.0.0.1 "IPv4 Loopback" dcs.log:28:2025-09-06 18:29:45.100 INFO ASYNCNET (Main): Adding LAN search interface 0: 255.255.255.255 "IPv4 Broadcast" dcs.log:356:2025-09-06 18:30:07.852 WARNING ASYNCNET (Main): Ignoring duplicate request to imageAFGHANISTAN_terrain while one is already in flight. dcs.log:357:2025-09-06 18:30:12.471 ERROR ASYNCNET (10472): Entering offline mode failed because server returned an invalid reply. Starting the game, authentication succeeds, but half a minute later, the server responds with an invalid reply. As per the OP, these are Error 500 (Internal Server Error) and Error 401 (Unauthorized). These HTTP error codes come from the server, after the client successfully completed an HTTP request/response transaction. Hence, no apparent network issue. That only leaves a server-side cause, so opening a ticket for ED support to get involved is IMHO the most promising approach.

I agree. It's a shame that adding an EWR to the group won't change a thing in vanilla DCS. Regardless, the higher detection range in a ~50° sector is nonsensical.

The SA-5 TR has a significantly higher range than its SR. With the TR emitting by default*, an SA-5 battery will lock and shoot a target within ±25° of the TR azimuth angle at ~130 nm range. This is far outside of the TR antenna main lobe width of less than 2°. When this happens, the TR azimuth will suddenly snap from its initial angle by 25° to the azimuth of the approaching aircraft: SA5_TR_Omniscient.trk The TR scans neither azimuth not elevation, sitting still until it magically becomes aware where to aim to lock the aircraft. When approaching 30° off the TR azimuth, it will only lock and engage an F-16 at a range of 70 nm. Presumably, getting the targeting information from its shorter range SR: SA_TR_Normal.trk *) Apart from the TR magically picking up targets outside of its antenna main lobe, I doubt it would be continuously emitting. Firstly, it would give away its location, inviting ARM attacks. Secondly, unlike an SR, the TR electronics would likely not be designed to emit continuously for hours/days, causing unnecessary wear and thus premature hardware failure. SA5_TR_Test.miz

When multiple players use dynamic spawn to spawn more or less simultaneously, they may spawn in the same parking position, resulting in multiple planes spawned inside of each other. This physically invalid state is then quickly resolved by means of pilot death, explosions, and magically levitating fireballs. This is not the first time, I've noticed this issue. Unfortunately, the associated track file is too large to upload (~80 MB). Instead, I've uploaded a fun explosion video, showing the spawn issue occurring twice in a row. First with two F-16s, then with three. Guess this is a server-side issue, with the server failing to arbitrate parking positions between simultaneously spawning players. dynamic_spawn_explosions.mp4

That was allegedly fixed according to the 2.9.19.13478 changelog: However, HARMs will also hit destroyed emitters with perfect accuracy. The attached .trk and Tacview files show an HTS-targeted PGM5 POS PB shot at an active P-19 followed by a second POS RUK shot. The second shot was fired shortly before the P19 was destroyed. However, it will still home on the dead P19: Tacview-20250901-193008-DCS.zip.acmiHARM_HTS_POS_hits_dead_P19.trk Same applies for an initial HTS POS shot followed by three HAS shots. All follow-up shots will home on the destroyed P19: Tacview-20250901-194932-DCS.zip.acmiHARM_HAS_hits_dead_P19.trk

@celestHawk Thank you for speaking out! I do agree with many of the points you make. Presumably, the silent majority simply accepts DCS as-is or – worse for us who stick around – simply drops DCS. Without a constant influx of new players/simmers, DCS will eventually die out for lack of funding. I'd love to see ED's statistics on retention of new users. Having started DCS two years ago, I can attest to the onboarding experience being subpar compared to other games/sims around. Without the community content, I would have been lost even more. But, as you pointed out, there's nothing like DCS, which is the only reason I stuck with it. DCS' bugs and issues puzzled me. I come from ArmA 1 thru 3 and DCS seemed like a noticeable step backwards in terms of documentation, scripting, and stability. The updater is archaic. ED promised background updates about a year ago but they never materialized. It should continue interrupted updates and also recognize insufficient disk space, though. Never had an issue with that. @Rudel_chw You are a decade-long, die-hard DCS fan. You are entitled to your opinion. However, sarcastically telling disagreeing users to just "bugger off" is against everyone's interest. DCS needs to retain and grow its user base to keep its business afloat. If a superior competitor ever comes around it may flip the switch on DCS. Having started DCS two years ago and having a software development background, I'm still unpleasantly surprised by its multitude of bugs and issues. While ED do solicit feedback and bug reports, I feel like the effort is in vain if such posts do not get acknowledged officially. In case you'd like to know more about my – certainly incomplete – technical assessment and suggestions for the DCS core, feel free to read this. No feedback from ED about it, just like several core bug reports, I've recently posted, which is frustrating. Also nothing new on the net.dostring_in() issue. An 8+ CVSS security vulnerability simply sitting around despite a mitigation being available. There's certainly room for improvement.

The DCS WebGUI app communicates with the DCS dedicated server via HTTP POST requests to /encryptedRequest. On the dedicated server, these POST requests are handled by a built-in HTTP server listening on :8088 (or whatever port configured via webgui_port in autoexec.cfg). This works well when accessing the WebGUI in a browser to communicate with a local dedicated server. Unfortunately, the URL is hard-coded into WebGUI/js/app.js as http://127.0.0.1:8088/encryptedRequest, which fails when providing access to the WebGUI via an HTTP reverse proxy. Supporting reverse proxy use would facilitate providing configurable, secure access to the WebGUI on dedicated servers without handing out the server account login credentials. This has come up in the forums a couple of times [1, 2, 3, 4]. Screenshot of the WebGUI issue with and unmodified WebGUI/js/app.js file: This issue is relatively straightforward to fix by removing the hard-coded URL from WebGUI/js/app.js and setting up the encryption key for non-local use with the following three search and replace operations: Search: dynamicUrl:"http://127.0.0.1\\:8088/encryptedRequest" Replace: dynamicUrl:"encryptedRequest" Search: "http://".concat(e.address,":").concat(e.port,"/encryptedRequest") Replace: window.location.protocol=="file:"?"http://".concat(e.address,":").concat(e.port,"/encryptedRequest"):"encryptedRequest" Search: if(e.url.includes("file:///")||e.url.includes("localhost")) Replace: if(true) Or alternatively via a single sed invocation: sed 's,dynamicUrl:"http://127.0.0.1\\\\:8088/encryptedRequest",dynamicUrl:"encryptedRequest",' app.js.orig \ | sed 's,if(e.url.includes("file:///")||e.url.includes("localhost")),if(true),' \ | sed 's;"http://".concat(e.address,":").concat(e.port,"/encryptedRequest");window.location.protocol==="file:"?"http://".concat(e.address,":").concat(e.port,"/encryptedRequest"):"encryptedRequest";' >app.js.fixed These minimal changes will enable use of the WebGUI via a reverse proxy, without affecting local WebGUI use (will work as before).

assert() is used for benign error handling. It is not a security measure against malicious code. The reverted changes attempted to address the latter.

Thanks! It does, but only with absolute paths (relative paths fail with error: 31) and string literals. That is not ideal, because the literals differ for every dedicated server instance. That requires to modify each config file instead of relying on auto-configuration. Attempting to dynamically set the writedir in options.lua via ["tacviewExportPath"] = lfs.writedir() .. "Tacview", fails for lack of the lfs module. Same with environment variables for lack of os.getenv(). Trying to set it from a hook script as follows net.dostring_in("config", [[ options.plugins.Tacview.tacviewExportPath = lfs.writedir() .. "Tacview" return options.plugins.Tacview.tacviewExportPath ]]) does set the configuration value, but Tacview fails to honor it. It uses the value set in options.lua. If you ever happen to have time on your hands, an option to save the .acmi files relative to lfs.writedir() would be greatly appreciated. Will do.

Totally worth a shot, IMHO. The Pentium only has 6% less single thread rating than my current Ryzen 9 5900X. DCS will only max out a single core and the OS scheduler should be able to manage that, unless its simultaneously transcoding some videos. Should achieve performance roughly comparable to my benchmarks with the 5900X.

Dummy response to mark as solved.

The CPU (Pentium Gold 8505) does have a fairly decent cpubenchmark.net single thread rating of 3272, which is only 36% less than a Core Ultra 9 285K. See CPU comparison. Should do fine for an average mission.

Anti-radiation missiles (ARMs) can lock on to and hit moving ships, but impacts only have limited effect on the health/hitpoints of the ship. Despite homing on radar emitters, ARM hits will not disable the radar in DCS World as of version 2.9.19.13478, because it lacks a ship damage model. This script adds the necessary event handler to permanently disable the radar of a ship that is hit by an ARM via :enableEmission(false). Works in multiplayer and on the dedicated server. Unfortunately, that will also disable all point-defense systems (CIWS) that should have independent targeting sensors, rendering ships entirely defenseless with the first ARM hit. While this is admittedly crude, IMHO, it's still a step up from DCS vanilla behavior of not modelling ARM hits on ships. Install via the mission editor: Open trigger screen (via menu item "Set rules for triggers") Triggers > New > Type: Mission Start Actions > New > Action: Do Script File > File: Open > ARMsDisableShipRadars.lua F-16 example mission (with debug messages enabled) attached: ARMs_Disable_Ship_Radars.miz Contents of attached ARMsDisableShipRadars.lua for a quick view:

Running Tacview with multiple DCS server instances on the same user account via different writedirs (DCS_server.exe -w DCS.server1..N) currently results in all Tacview files from all instances being combined into a single folder. Is there a way to have Tacview store the files within the profile/instance/writedir folder, i.e., %USERPROFILE%/Saved Games/DCS.serverN/? The only configuration option I found so far is the following registry key: [HKCU\\SOFTWARE\\Raia Software Inc.] "TacviewExportPath"="C:\\\\DCS_configs\\\\Tacview\\\\" Being part of HKCU, it results in a combined folder for all instances/writedirs as described above. What do I need this for? I maintain a project to facilitate running the DCS dedicated server on Linux via Wine. It supports multiple server instances with separate writedirs. After manual installation, Tacview runs well with it and I'd like to streamline the admin experience by separating Tacview files per server instance to provide fine-grained remote file access on a per-instance basis. Setting the directory via environment variables or alternatively a variable in the Tacview scripts or autoexec.cfg would be an option for me. Basically, any way to do this in an automated fashion would suit me. Also, to facilitate server-side Tacview deployment, I'd like to install the Tacview hook scripts and DLLs automatically. Would you be open to that? It would need a straightforward access to the relevant files, short of downloading the ~1 GB installer exe and extracting the required files from it. Feel free to DM me on this matter to discuss details.

Still bugged as of 2.9.19.13478.