Notice Fleet Carrier Update - Known Issues

yeah... Black Adder
tried in a few different systems :(

Update: just tried CD-47 990 as there is just 16 FCs in the system, Black Adder in any game mode... what a heck?

There is a webservice that the game calls out to to coordinate the positioning of FCs. The bug is that for some reason we all end up asking that webservice for the same two FC ids. It doesn't seem to have anything to do with what system you are in or where you are. Some bug is providing the wrong next ID for new carriers and gives IDs that already exist resulting in the registration service returning a failure. That failure is then not well handled by exception code in the client (bad programming) and the game client exits to protect the game state against an unexpected condition.

That's what I've been able to determine as an unaffiliated SAAS architect/developer with over a decade of software development experience. I don't see this issue getting fixed until FDev has a chance to debug it during the upcoming workday, but expect they'll be able to find the problem (which is probably something relatively simple) and then everything should start working again.

The missing locations is of bigger concern to me as that's likely significantly harder to fix and likely has something to do with procedural generation ending up giving the fixed POIs the wrong coordinates now. Who knows how that remap will work or how much effort it may require.
 
since your so wonderful, how about you develop, code, model, texture, uv map a better version for yourself... so sick of peons complaining all the time when they have no idea about the amount of work that goes into a gametitle
Every single big patch day for ED has been like this, and there has been MUCH worse, and yet people still act surprised. I still remember the original Engineers update where I couldn't even log in for 3 days
 
The missing locations is of bigger concern to me as that's likely significantly harder to fix and likely has something to do with procedural generation ending up giving the fixed POIs the wrong coordinates now. Who knows how that remap will work or how much effort it may require.

I'm not sure how source control integrates with their various build. And I think you are right.

So why don't they just roll back the patch?
 
it might have to do something with the procedural generation doesn't allow for a realistic rollback path. though i'm no expert on this stuff :)
 
Anyone know what "LOGIN FAILURE, Connection Status: Invalid machine ID" means? Frontier are asleep despite this being a global game.
Never happened before the latest update, verify files didn't work.

Capture.PNG
 
Last edited:
I'm on PS4 and I seem to have no trouble logging into and playing the game but as soon as I try purchasing a FC I get the 'Black Adder' something's gone wrong message and sent back to main menu, I've tried every half hour for the last 4 hours but it's the same thing every time. That's enough punishment for me until tomorrow.
 
since your so wonderful, how about you develop, code, model, texture, uv map a better version for yourself... so sick of peons complaining all the time when they have no idea about the amount of work that goes into a gametitle

Some of us do develop our own products. I personally am an architect, framework developer and build/version control manager for a product that ships 3-4 major updates a year with similar levels of staffing to what Frontier most likely has and do so with 99.99% up time and zero critical bugs most years (and generally under 10 high severity bugs). I do agree that some people have unrealistic expectations, but if this release is anything like a typical FDev release, it leaves a lot to be desired.

I've spent the last few hours investigating the black adder bug to see what is happening under the hood and while I actually expect the underlying bug is probably relatively reasonable to have occurred in some edge case state they didn't expect, the lack of error handling around the particular call (which appears highly isolated from the rest of the game state) is just bad software development practice. They are literally just making a webservice call out to an external coordinator microservice that appears to provide FC positioning and status data to the game clients. They are feeding it one of two bad new carrier IDs depending on some unknown condition that always produce one of the two same IDs. Since these carrier IDs already exist, the microservice returns an error that the carrier can't be spawned. Rather than gracefully handle a failure from an external web service independent of the main game state, the code then fails to the menu. That's the kind of thing I would notice in a code review and make a developer fix before it even got to testing.

Don't get me wrong, I love FDev. I love Elite Dangerous. I realize every developer has their relative strengths and weaknesses, but that doesn't mean they don't have a lot of room to improve and really should focus on trying to improve their code quality. (Note, I'm a software architect so I'm of course biased towards improving code quality as it's a significant portion of my job to come in and help fix up bad code to make it less bad code and eventually good code.)
 
The bug is that for some reason we all end up asking that webservice for the same two FC ids.

Thanks, just not completely sure if you are right about FC ID. I realized something interesting: Every single time I bought the FC (Beta1, Beta2, and trying now) the vessel ID was always the same (V08-Z4L) as it was pre-generated per commander. I just hope they fix it soon. Reading all the news after a busy day I cannot agree more with you: other "ripples" are scarier than the actual inability of purchasing the FC
 
To err is human, to really screw things up you need a computer.
I have worked in IT and I understand that patched can go south really fast.
I wish the team luck at finding the bugs and crushing them.
Please don't take to long. I recently introduced my two oldest children to Elite and I want them to enjoy it.
Kind of hard to do right now.

I expect the FC buying bug will be quick if they prioritize it. Something is just giving out invalid IDs (and consistently the same invalid IDs, so it should be really easy to troubleshoot). I'm more worried about the missing POIs, that could potentially be problematic depending on how they have it architected and how they overlay the POIs on the procedural world and based on what exactly went wrong there. (It might be easy too, depends on if it's an addressing issue or a layering/application of POI issue.)
 
Thanks, just not completely sure if you are right about FC ID. I realized something interesting: Every single time I bought the FC (Beta1, Beta2, and trying now) the vessel ID was always the same (V08-Z4L) as it was pre-generated per commander. I just hope they fix it soon. Reading all the news after a busy day I cannot agree more with you: other "ripples" are scarier than the actual inability of purchasing the FC

This is not the same ID as the number. The IDs are 1600684 and 1601112 and have been reported by multiple users on the forums as being the IDs in their verbose logs. For the curious, the details are below. Note that I removed two large encoded bits of the log that may or may not contain private or secure data given that I was not able to figure out their encoding scheme:


{00:02:44GMT 184.204s} Webserver request failed: code 403, details 'removed from post by me'
{00:02:44GMT 184.204s} Webserver request: https://api.orerve.net:443/2.0/elite/fleetcarrier/vendor/buy?removed from post by me...
Plaintext error response: {"status":403,"message":"Helper\\FleetCarrier::spawnCarrier Failed to spawn Fleet Carrier 1601112 because we are already spawned"}
 
Seems to be a sound bug prevalent where sound affects from the hangar are still prevalent while out in space. my ship has all kinds of static sounds and industrial machinery sounds buzzing all about even in deep space. it also now sounds like my ship keeps hitting rocks while cruising around. hope this isnt intended as this is pretty annoying after only 5 minutes of playing
 
The missing locations is of bigger concern to me as that's likely significantly harder to fix and likely has something to do with procedural generation ending up giving the fixed POIs the wrong coordinates now. Who knows how that remap will work or how much effort it may require.

I am definitely concerned about this too, all the other stuff like actually buying a carrier and system POI's acting weird when carriers are around are the kinds of things I expect when an update goes live at this point and I don't care that much.
 
I'm not sure how source control integrates with their various build. And I think you are right.

So why don't they just roll back the patch?

They may not have roll back scrips for the data schema. They could do a revert, but then they would lose any progress for anyone that was doing stuff today. If they have a reasonable handle on the problems, then it probably makes more sense to charge forward than roll back. You also would have to coordinate multiple deployments of the rolled back version across multiple platforms. It's likely non-trivial. I'm not at all surprised they didn't just go with a roll back or a revert.
 
Back
Top Bottom