Samoht

During the course of a multiplayer mission, the carrier TACAN will frequently go out of service. It may possibly be related to people hitting the back of the ship with failed landing attempts, but I can't confirm this. Ships in the mission are set to immortal.

When setting all available equipment aboard a carrier to 0, in multiplayer, equipment is still treated as unlimited. For non-combat servers like ours or missions where certain types of missiles or bombs are not meant to be available, this is game breaking.

Several bugs exist with this: * Pylon state is not synced in mutliplayer. A Blue Angels aircraft with pylons removed will render properly in the client's game, but will show pylons for any other multiplayer clients. * Pylons are automatically readded to the client when performing any action in the Refuel and Rearm menu, regardless of adding payloads or not. * It is not possible to remove pylons after they have been added. Related: aircraft equipped with pylons at spawn cannot remove pylons.

The stability and network performance with Update 3 seemed to have been the best I've ever seen it in DCS. That was undone entirely with Wednesday's Update 4. We're seeing many more instances of lag spikes and warping aircraft in the sky, but especially on the ground. Several player jets are hopping 200 feet into the air over and over. Many are being caught by our taxiway speed enforcement scripts even while they're not moving, as the server thinks they are. There is no change to the server or Windows installation aside from the new patch being installed.

There does not seem to be any change with stability regarding the Harrier in this latest update. Allowing people to fly it in multiplayer chokes the server, makes people warp, and eventually freezes DCS (but not crash). I've once again removed it from the mission to make the game playable.

It might be the ED listing server resetting and losing your server. What is your server's restart cadence?

Yeah, I noticed the parking spaces all adjusted. I spotted that one in the first 2.5.0 OpenBeta and got that sorted out. We don't guard against people connecting on load, but then again, it's never been an issue for us (I don't know why, since it obviously is a problem for others). The crashes we're getting on load don't seem to be related at all to people connecting. We're seeing this: 2018-05-08 02:08:13.169 INFO EDCORE: C:\Program Files\Eagle Dynamics\DCS World\bin\World.dll 2018-05-08 02:08:13.169 INFO EDCORE: # C0000005 ACCESS_VIOLATION at 5A120700 00:00000000 2018-05-08 02:08:13.172 INFO EDCORE: 00000000 00000000 0000:00000000 2018-05-08 02:08:13.173 INFO EDCORE: 5A120700 6716DF10 0000:00000000 C:\Program Files\Eagle Dynamics\DCS World\bin\World.dll 2018-05-08 02:08:13.174 INFO EDCORE: 58F50848 6716E1E0 0000:00000000 C:\Program Files\Eagle Dynamics\DCS World\bin\AsyncNet.dll It'll rarely load clean without a crash. Usually it's once. Sometimes it's twice. The process becomes non-responsive and the external watchdog will kill DCS and restart it after a few minutes.

I'm not sure if ED runs or has ever considered running a proper multiplayer server that users will regularly use, but I think it would be beneficial if they did. That way they can see first hand the issues that we constantly deal with and have direct access to logs and crash reports.

I can also confirm the Harrier causes major issues with Aerobatics Online. It took awhile to narrow down, but after pulling out all Harrier slots, the server warping and UI freeze issues have completely gone away.

Rico: Surprisingly, we're not really seeing the spawn error that you've been getting. If it has happened, it's infrequent enough that I don't notice. I've pulled the Harrier from our mission and doing so has drastically helped with warping issues. I've done several long flights when the server is full and not had any warps. Crashes are still happening though. Mainly on initial start. The DCS process crashes and then hangs when first loading. My watchdog script will get it eventually, but it delays server availability after scheduled restarts. Slick: I rebuild the AO mission from scratch about once a year, but I haven't since 2.5 came out. With every new update, at minimum I save the mission in the new version and fix the airport inventories, but redoing Caucasus from the ground up is about three days of work. I should probably do it again for 2.5, but I'll wait for this current version to migrate to stable before undertaking that. Tovivan: The mission editor is extremely unfriendly with customizing weather and getting it where you want it. It's pretty much random chance that you get something that works. In our case, I need to make sure that the winds are low enough at spawn airports but challenging enough at other distant airports. When we have racing going on, I also don't want the moutain courses to be covered in clouds. It's almost impossible to get all the weather to cooperate in all locations that matter. That being said, we need to come up with some new activities and challenges. Doing more with the weather is high on my list. This is assuming that the weather bugs that got introduced with 2.2 have been sorted out. I haven't checked.

Still having on average 9 crashes per day. When it's running, it's pretty unplayable (for formation flying). It's been a long time since the server has been this empty for this long. I'm continuing to reduce mission assets, but there doesn't seem to be any significant change with stability.

We've been getting multiple crashes and UI freezes per day on the Aerobatics Online Caucasus server ever since the latest OpenBeta update. As other admins have pointed out, 2.5.0 was not doing this, but 2.5.1 is crashing a lot. Unfortunately, we don't get any crash logs when the entire process freezes; however, an outright crash will generate a dump file. A dump file is attached. We're powering through keeping and OpenBeta server up to hopefully provide some good data you can use for making the next patch better. In the mean time, I spun up a second AO Caucasus server on DCS World release. They're running pretty much the same mission, and the computers hosting have the exact same specs. It might make for a good comparison. As a side note, aircraft seem to "hop" now when flying in close formation. It's not always apparent when observing someone far away, but in close proximity (less than six feet) the judder makes things nearly unplayable. It's not the world, just the other aircraft. We did not see this when flying on a small mission and it's not always present on AO either. I also have not seen any significant reduction with mission load-in time when using a large preload radius, which is necessary to cache models and avoid spawn related lag spikes. dcs.log-20180428-033349.zip

Per server, on the weekends, it's usually in the 30s, sometimes going into the 40s. During the week, it's in the 10-20s. Since swapping the video card from a 1050 2gb to a 1050 ti 4gb, it doesn't seem to happen as much, but it still does occasionally. I don't know what the root cause is, but nothing I've done to scripts or the mission has changed things as best as I can tell. Edit: this wasn't an issue until 1.5.8 and 2.2.0.

Yup, that's pretty much exactly what I see on both of my servers too.

I captured the logs, track, and Tacview of the particular instance I was talking about. To capture this exploit via scripting would require a lot short interval polling of all players since this exploit is so fast. It's something to investigate, but for now I'm running integrity check. ED: if you want the full logs, I will share them privately.

Since it's an aerobatics server, all of the airport inventories have been cleared of munitions. It was my understanding that, even though you can modify your aircraft's loadout, you still need to have access to weapons (which aren't available on this server). From the client side, it's still possible to load a nuke? That said, the method of detonation is always the same: go straight up, then drop it down like a meteor.

I've been having problems with a small number of players taking advantage of some exploit in multiplayer, allowing them to somehow slew their aircraft to extremely high altitudes, and then drop it to the airport at very high speed. The impact causes what is effectively a nuclear explosion and destroys the airbase. It appears this can be done with any airplane with no weapons equipped. Does anyone know how they're doing this, or more importantly, how to stop it (short of using integrity check on the server). This seems like a bug, as I can't think of any use case where this should be allowed in multiplayer.

I run and maintain the Aerobatics Online servers. Our 1.5.8 release server has been freezing multiple times per day. This problem has existed since 1.5.8 was initially released on Open Beta and was brought over to Release with the last update. I managed to finally catch it in action and manually completed capturing the logs. Normally it just hangs and my automation scripts kill the process before it finishes getting/sending logs (because they don't send automatically, and they should). Symptoms: After some random period of time (could be 20 minutes, could be 120 minutes) connected players will begin to warp around. On the server's console side, the interface gets intermittent freezes (blue busy mouse indicator) or periodically switches to the black DCS logo screen. If I keep trying to interact with the console, Windows will eventually report that the DCS process is not responding. Eventually, my watchdog timer will fire noting that the simulation frames have stopped and kill the DCS process. If this is not done, the DCS process will hang forever (or until manual intervention). Players are usually not disconnected when this happens, but they are frozen in the sky. There does not seem to be any particular trigger as to what causes this. The attached logs were when the server only had three people on, but it was after crashing once previously. Most times it crashes with 15+ players connected. AO Caucasus is running a stripped down version of the mission with basic assets only. The only AI units are tankers. Ships are statically placed at runtime using external scripts. Other administrative functionality is handled through external scripts that are not included in the mission. I have tried running the server with the scripting interface down and still get similar performance/stability issues. Server specs: Windows 10 version 1709, i7 7600 (stock), 16GB ram, 1050ti 4gb video card (stock), SSD with 42gb free. This is a computer dedicated to running DCS World as a server. The computer automatically reboots after every game session (4 hours, currently) and does have a sound device enabled. I should note that the DCS World multiplayer server hardware requirements have dramatically increased with this latest version. More than ever, the video card makes a difference whereas it just needed to support DirectX 11 before and a GT 720 would happily run a multiplayer server. That's no longer the case, even with options.graphics.render3D = false in the autoexec.cfg. These stability issues are killing our player base. On a Saturday at mid-day in the US, especially right before Christmas, I should have 45 people online right now. Instead, I have eight. dcs.log-20171223-203347.zip

It's good that you might have found a root cause to this problem. Hosting Open Beta for Aerobatics Online, I ran into the exact same issues you're describing. I run on several 2U headless servers, none of which have sound cards, but all have low power NVidia video cards. I was completely unsuccessful in getting Open Beta to run without long hangs. I thought it might be GPU related, but maybe it's not...?

Yup, you're right. I patched Nevada again from a different computer and accidently overwrote the mission options with old settings. It's fixed again now.

Aerobatics Online Nevada and Normandy has tooltips disabled with regard to this bug.

The Aerobatics Online Nevada server is running in a new home. Since the machine and IP address have changed, you'll need to add it to your favorites again as it'll not be starred anymore. Tell us if you see a difference, especially with regards to lag spikes. We're interested to know!

All of the Aerobatics Online servers require simulation controls, so that may be your issue. I'm not aware of multiplayer missions being able to modify or corrupt a player's controls. If that's happening, I would suggest reporting that to ED. I'm not sure which AO server in particular you were referring to regarding 30 second lag spikes. I've only very rarely gotten lag spikes that were longer than half a second, personally. A longer spike is usually indicative of an impending DCS server crash. Our Nevada server has been suffering multiple freezes per day, so if that's where you're playing, that might be the explanation. As to a long term solution: we're exploring options with more physical server hosts to host maps. I've temporarily taken the Normandy server offline to see if it was conflicting with the Nevada server since they run on the same physical computer. We'll make announcements when any major changes happen. Also, I stole your forum signature, Nelson. I didn't have one and yours seemed nice.

Seconded. I would like the ability to disable this if necessary for multiplayer servers.

Looking at server logs, we're seeing around five to seven crashes a day during busy periods. This is still a big issue for us.