@lancake apolgies for the late response, but i don't believe i have the log file anymore and if i do i'd have to dig deep to see if it is still on record.

They do not have any permission module beside the rail.

I tried it for a few more time and find out that power would only drained if you are playing in an existing universe prior to the patch and loaded in the sector for the 1st time.
If you leave the sector unloaded for a while, the power doesnt drained anymore.

In T2524#107136, @AndyP wrote:

Can you upload the logstarmade.0.log from the server-side from a broken startup?
There should be some error, that locks parts of the server.

@AndyP - Not surprised by your diagnoses. I will look for that log file.

happened again, game started crashing after update. Logged back in later and faction/credits/saved nav points are gone. I can fix faction, faction points, credits but nav points are toast.

Iancake, Herald is referring to the intel graphics glitch which is incompatible with frame-buffer. To fix the annoying blue shade turn framebuffer off off.

Bug no longer appears to be present in the latest release, beams are doing propper block damage cross sector in all cases I can think of to test.

We have the same problem and I have took that from our server log:

Im getting a LOT of these recently now that Ive started back into fighting other ships and testing. Retry is a temporary fix if it doesnt outright crash.

i logged onto the server i play on after playing on sp and i was almost constantly getting the error (we where under attack) so here is a new set of logs. also cannon shots and missiles froze in mid air after and during the error.

I was able to reproduce it using the process above on a new world, this time with missiles and cannons/cannon/explosive effect. I am uploading the world and logs, one of the ships i was able to reproduce it with is in sector 50, 500, 500, the turrets are cannon turrets.

Old bug that was tackled in another task, might still happen though but that will be looked at in the other task.

Awesome! Found an existing of the break off issue but it's a bit outdated. Will merge the tasks together and update the original one.

Yes, disabling break off appears to fix this issue, thanks for the feedback :)

Apologies on the delay there

Going to main menu and back in-game several times in the same session will indeed result in a memory build up, although that should steadily decline after a while but usually at a too slow pace to help.
Frequently alt tabbing can make the GC go lazy too often and could also take a while for it to recover completely.

I see. Main issue is then that the "Set price to Buy" and "Set price to Sell" for the shop owner (only the shop owner gets this anyway), should be switched around when it comes to function.
We might as well give them the exact same name as in the Trade Order one

Yeah, can reproduce fine now, probably messed up and forgot that '+ protected' isn't persistent between shutdowns.

Interesting details, missile targeting changed quite a bit but this could give us clues of what's happening in another issue. Merging them together.

Merging this issue into another one.

Well I can reproduce some of it, but doesn't sound as severe as you describe it.
Basically anything from a single tick can only apply either planet core damage, or block damage.

Should have been caused by a NaN exception upon undocking. Shootout rails pointing towards each other and including pick up rails close by would have triggered that.
That's fixed, or at least most cases are fixed...any remaining issue will be handled in another task. Merging them together now

Server Log file:

In T2328#101765, @lancake wrote:

-feedback-
Unfortunately I can't reproduce. I have tried it out with a quite large sector (10 mil blocks total), even making sure the spawn location would be occupied by chunks, loads in just fine and time to load/spawn is pretty much the same as usual.
Do you manage to get past the loading screen and you just end up unable to spawn after clicking the spawn button, or do you timeout during the loading screen? Could you upload your log file too for the client that is timing out? Logs are located in the Starmade/logs folder, logstarmade.0.log is the most recent one. Upload that file here by drag and dropping it in phabricator's message box.

happens almost every time i use missiles on multiplayer servers

I can second that we are getting this problem on the Brierie server too. But we also have a problem with loading entities in shipyards (as detailed in T2408)

I can say Ive experienced this

Here's the shipyard attached. If for whatever reason you cannot replicate it in a multiplayer enviroment, I assure you just about any shipyard on the Starmade Australia server has these issues and I can provide video footage if need be.

On more testing it looks more like editing a ship is what changes the tracking. The first time I put a missile computer down the missile seems to track but if I delete it and re-place it it won't track properly so width doesn't seem to be a factor.

As above, its Upload from Local to Multiplayer.
Also wish we can search in that menu than have to scroll a very long drop-down list...

Probably the local blueprint upload list dialog box. Upon pressing that, I get a very long program hang myself, so I generally hit the button and then alt-tab out and do something else :P

In T2385#100190, @Ithirahad wrote:

It is a single activator triggering a large number of forcefield and forcefield wedge blocks distributed across several chunks, likely more than 50000. But this, alone, should not qualify as a 'logic bomb.' It doesn't alternate or repeat automatically to produce a sustained load on the server. Essentially, this 'feature' is just punishing large-scale implementation of the logic system that does not produce a noticeable effect on servers, which I can't imagine is an intended result. (therefore, bug.)

Im having the same issue. My setup was a plexdoor box that closed around the entity while it was still on the rail, then a shootout rail to disconnect, and a pickup rail on the receiving entity. It all works as intended, The plexdoor enclosure keeps the entity in place when the shootout rail disconnects it, then after five seconds the pickup rail attaches it the next entity. The problem is that at that point, the object im moving is now stuck. Even breaking the rail docker wont free it, Only after I restart my game will it unfreeze.

I have had this issue when using select target lockon missile turrets

The game hangs for 7-10 seconds when I open the menu, but that is significantly better than crashing outright. Somewhere its fixed itself.

My homing missles do all diverge from the locked target.
In single- and multiplayer. They diverge from the target in perfect curve away from it into empty space.

Just logged back into the multiplayer server and all Rails have lost their links to all Stores on my station, and on all ships docked to my station. So contrary to what I previously witnessed, its not a couple of Rails, its all of them

Thanks for having a look lancake - Id previously thought it was that issue, so I tested it myself (the first time it happened was on a fleet miner carrier), but I couldnt reproduce it like that either.

even now I still produce unmovable cores. But only on the server. In a local game there s no problem

It is a single activator triggering a large number of forcefield and forcefield wedge blocks distributed across several chunks, likely more than 50000. But this, alone, should not qualify as a 'logic bomb.' It doesn't alternate or repeat automatically to produce a sustained load on the server. Essentially, this 'feature' is just punishing large-scale implementation of the logic system that does not produce a noticeable effect on servers, which I can't imagine is an intended result. (therefore, bug.)

We're taking some preemptive action tomorrow to clear out the blueprints on the server as we're starting to get instability indicative of the type caused by this bug.

Yes, with double quote (" ") only for the "command" param, and the escape sequence :)
Thanks and Sorry for forgetting to report it.

We have some additional information for you.

Another player claims to be having the same issue on the forum. They posted in the thread I started, so this might not just be limited to me.

Please see attached.

I have just tested it, and it teleports to the correct position promptly. When I reported this I believe the rail entity was stuck and teleported to the original position, but it seems to be okay now.

Yeah, that was a mass-comment to usually poke those people again that did not reply on our feedback request, looks like your reply slipped through. (And this was the exact reason we switched toward auto-tagging, so replies do not rely on reading notification mails or looking for them by hand.)

Waiting on Author?

Oh, yes that could indeed be a factor. I tried them vs real enemy factions, not the neutral=enemy setup. Will try them again with this mode.

Yes I did all those steps and I always check the ai module after entering/leaving the core to make sure the settings are correct.

The turret is on an elevated flat platform on my station so im not sure what it possibly could be running into. You can pilot it just fine and it can turn in all directions other than straight down.

Nevermind, it's already been fixed in the pre-builds. Most likely related to T1299
Don't need to set this one on private though, merging tasks.

@AndyP Oh, this is old. It hasn't happened since then. The "textures not loaded" bug has also since been fixed (Though it still seems to happen sometimes when quick switching between servers, without fully closing the client and re-opening it).

Yeah, metadata is kept in place, even if the display is removed, kind of another bug, but unrelated to this.

I tried removing the old MASTER diplay and build new ones in the same spot and they automatically have the same text
Had an similar bug with Rails when i connected them to a button removed the button place it again (same spot) and it still had the connection (C button V Rail)

Oh, okay, I understand.
should be enough to test it.

I found the problem the day I posted the report and the former command without the "24m x 15m" worker for at least 1-2 months

I understand what Jeryia is saying. If there in no faction block on the ship anyone can just take the ship, you need to place the faction block add you faction then set it to owner, but you can only do this when in a faction. so basically if you are not in a faction you can not use the faction block and in turn you can not lock your ship. yes you can make a faction of 1 person and then do this to lock your ships but if you don't want a faction ect ect, then you can not lock your ships. I just tried it with my brother. if I build a ship and then exit the ship he can just take it, if I use a faction block then he can not take it. its pretty simple really

I doubt 'we' can do much about this, but this is for sure something that may be tweaked at a later time.
Like always compress inside java and write a sparse file compatible format, and use the same way to extract, no external tool requirement to restore a backup can help with this.

Perhaps this is better phrased as "windows backups need sparse file support", as tar/gzip support was added to the linux side a while back.

This problem have been resolved by the latest update. Didnt have anymore problems in saving and spawning a blueprint.

Quoted from erthparadine:

side-note: that Intel GPU warning is useless for macOS users…at-best its an annoyance that nobody can do anything about. At worst, it suggests SM developers are not aware of the inherent driver integration on the macOS platform, and therefore that nobody can simply “update” drivers on macOS.

the server I'm on is having the same problem, its even corrupting the factions because of this issue. we've seen this bug since the faction update with npcs.

The set-up explained in the original description does not replicate it anymore, however I still do get the error but it is very hard to reproduce it reliably (however when using a missile system that causes it, it happens very reliably), I have only experienced it twice since then, both in multiplayer. I am using a Nvidia card as well. However, another SM player has this issue on AMD cards as well.

I'm pretty sure I have triggered this myself on many different public servers unknowingly. Right now I'm playing on Brierie (running .472) and I attempted to "move" my shop to another physical position on my station. Pretty sure it caused the server to need to hang/be restarted. But post restart things are still acting quite wonky at least in my sector.

Yeah, better to upload those to be sure although I should be able to reproduce with this sector crossing part.

This is what I see, while logged-in as a normal user, about 2 minutes after running /add_admin against my user.

@lancake I repeated the test as requested, and sure enough with both ships in the same sector I can damage blocks with the beam weapon as normal immediately after spawning. However, If I train on a second target (same blueprint, same firing ship) in a different sector, it exhibits the problem I was describing.

I have tested this bug out and found where the issue lies:
If you have a drone docked to a mother ship on a rail, then change out the rail it is docked on for a shoot-out rail the invalid position error occurs.
A workaround for this is to only swap out the direction of the rail your ship is docked to and have it lead to a shoot-out rail.
I was able to re-produce this bug with the supplied blue-print and the workaround consistently solves the issue.

Is there any update on the progress of this bug?

This still occurs, just try to run any tutorial on macOS, and the same crash happens.

Still seeing this issue. I suspect it's related to our application of the /add_admin and /remove_admin commands.

@lancake asked that I upload our factions.fac file. It is attached.