SlippyCheeze.5483

Interestingly, I found that there is at least one checkpoint in the LS4 E3 story, thanks to accidentally double-clicking the wrong thing and teleporting out of the instance.

Yes. For the players who suffer this problem, it happens with a very high probability at any cutscene. Since those only occur during missions... FWIW, I don't think this is the "players fault", but I do think that the likely cause -- which ANet are still trying to find, and have not yet succeeded -- for the small number of players who this hits is going to turn out to be outside their servers. Unfortunately, I had half hoped the migration to AWS would fix it, because that changes many of the potential root causes at their end, but it did not. Anyway, if you don't mind: what model and brand of router are you using, and are you using a wifi connection to play?

Oh, interesting! I hadn't considered that, and that is a useful point. I'll ask about that, too, and see if that turns up a commonality. The time does seem to be ~ 20 seconds idle to a disconnect, pretty consistently from the tcpdump traces, but ... if some do that when the link is idle for a minute, I can imagine someone cutting that down further because power saving reasons or something.

Is this something you might consider extending in the future? I ask, because I'm a pretty casual WvW player, honestly, and I'm currently on a T1 world. Everyone else at that level is way more serious than I ever expect to be, so I'm a bit of both an easy target, and inclined to feel like a bit of a hinderance -- even if I play competently with the groups, I'm likely to make unwise decisions because I just don't have the same commitment they do. From my perspective, the biggest hope is that I end up on a more "casual" sort of WvW environment. I'm not looking to suddenly be great, but I am hoping I might end up being less standard deviations from the usual effort put in by other players in the same grouping. Even just using "number of hours played" as a proxy for how much WvW someone does, and so preferencing groups so they mostly have similar sorts of hours-per-week counts, would be a big improvement to my chances of getting WvW that I enjoy more, I think. Either way, I'm looking forward to the new system - and not just because I'm too cheap to buy the gems and transfer to a less serious WvW world. :)

I'm not at home right now, but I'm with Xfinity. It's just really odd that it happens so often with Guild Wars 2, but doesn't happen in any other online game I play. Drat. That kind of shoots my idea about maybe a router commonality out of the water, because that isn't a VirginMedia Hub of any sort, which the first few were. During the cutscene, there is no network traffic between GW2 client and server, and at about the 20 second mark is where most people who suffer this disconnect see it happen. Most online games never go silent for that long, so that is the single most obvious difference between the two. Don't get me wrong, too: I'm not saying this isn't a GW2 issue or something, and I'm definitely not opposed to things like checkpoints to reduce the pain; I'd even encourage them to add a keepalive message of some sort during that period so the connection isn't completely silence. I'm just hoping that maybe we find a commonality, and then they can fix it right -- because even with checkpoints, if you disconnect seven times in a row it will suck, just suck a little less than it did. Anyway, yeah, GW2 clearly has some sort of issue here, but I couldn't for the life of me tell you what it is, and sadly, ANet have confirmed -- via posts from the developer who is actively working on this -- that they cannot yet determine it either. Kicking through to my next vague idea, do you have any sort of "internet security" software installed? I'm thinking things like the Norton Internet Security system, that sort of thing.

You know they don't actually ignore these, right? They have people actively working on them, they fix issues when they can, they just don't pop in to the forums and comment on every single report of disconnects, high latency, etc. (Not to mention that a bunch of them are beyond the control of ANet, since you get at least a 50 percent hit rate on that when people actually bother to do diagnostics about where on the network the issue is.)

+1 on the "small number of players" having this problem. To anyone reading, don't get me wrong: it super-sucks for the people it happens to, and it is annoyingly frequent for them. ANet have a developer tasked with working on figuring this out, but by their comments, they don't yet have any common pattern in what is going on, so no solution in sight quite yet.

This is a persistent problem for some players, and anet are looking into it. Could I ask, please, what brand and model of router you have?

The design is that you open all ten, completing the Master Diver achievement, and then they all reset so that you can do it again. This may not be what I would have preferred, but it is how it was designed, and that is unlikely to change regardless of the bug.

If you can, immediately after reset is one of the most active times on the US servers for that map; you are most likely to get a "fresh" map instance then. Clearly this is not a solution, but it is a work-around that got me my Subject S and, thus, backpack. Until ANet fix the underlying issue, hopefully this is sufficient. Still bugging (for me, at least) as of today (June 29). Event progressed all the way to the fight with the security golem(s), which were successfully defeated - then just sat there.Dang, you got to trigger the bug, not just run into it. That bites. The good news is every patch brings a new reset, so a new chance. The bad news is, whatever causes these stalls is apparently quite difficult to find and/or fix, so they can persist for vastly longer than I would have expected.

Still have to complete the achievement by opening all ten chests before they reset. Were those ones already completed in Master Diver?

When you dismount there is a little delay as they resummon / respawn automatically. As near as I can tell, if you mount before that has "completed" you lose them, I guess because whatever test says "if minions, despawn them and remember this" looks, and sees "no minions", or something. Pausing just a fraction longer when using a minion build made the issue go away for me, and it has only ever happened when rapidly switching mounts.

Great advice. Like "check the network", this is a very low cost thing to do, and ensures that you have a fully functional and correct client. You pretty much can't lose: it either fixes the issue, or you lost a couple minutes and are no worse off that you were before.

It seems this is purely about like and dislike. I think most comments reflect that. Mostly the reason people use "to give the story impact" or "to really draw us in" are just an attempt to cover it up. If I like (sometimes even hate) a character, I'm more apt to dislike them dying because creating characters I respond to at all is seemingly difficult. Especially if they're given any type of cohesive personality. I can think of some characters that just have no affect on me at all and its difficult to care at all what happens to them. I try. I'd rather have the hatred of a character. This. I don't like Braham much, and while I completely disagree with some of the views folks have of him, and how you should respond to his behaviour, I can understand where they are coming from, and why the would, for example, see his actions as so disrespectful that -- for the murder hobos we are in-game -- we should kill him. (IRL, just ditch him forever would be the equivalent.) That usually ends up in "so just kill him plz" though, because of what thruine says here.

I would love seeing a mix between a Sylvari and an Asura, would be funny to see a little walking cabbage I would play the compost out of that. Just sayin'

Hey, y'all. Please be aware that the disconnect issues in this thread and distinct from the issues around the launch of this patch. If your problems only started with LS4 Ep 3, a new thread (or just being patient, as ANet work on it) is the appropriate place. The thread is about a problem that hits some players, causing disconnects unrelated to server issues, unlike the current pain.

Was the Timberline Falls one already "completed" in this round of the achievement? You still have to unlock all ten and complete the cycle before they reset after this fix. It didn't reset progress on Master Diver.

They issue has been fixed?!? Why was there no update note??? Not sure, maybe to be sneaky? Or to keep all the shines to themselves? Elsewhere, the truth leaked out... some developer was a teensy bit lazy about sending in their contribution for the patch notes, and it missed the deadline. (Presumably for translation to other languages, etc) (I assume a developer, because as one, I can tell you we are super bad at doing paperwork in a timely fashion.)

I'm curious, what brand and model of router do you have? VirginMedia SmartHub 2.0 i think AFAIK.Interesting. That makes two out of two people with this problem, and that router. Not a very convincing sample yet, but an interesting coincidence. I strongly suspect that, unless you refuse to download anything in the launcher, and instantly jump into the mission, it will not end up helping. I'm totally with the idea of checkpoints, to be clear.

I'm sympathetic, and I know the question was kind of rhetorical, but ... yes, it really can be that tricky. Outwardly simple issues can have annoyingly deep causes in software, and worse, resolving them can be challenging without breaking something else entirely -- especially when your code covers a decade or so of regrettable life choices by past developers. One clue that a bug is going to be annoying, subtle, and slow to fix is how it triggers. In this case it was triggering unpredictably, part way through an achievement cycle, for only some players. That suggests that the trigger conditions are not simple, and you have to hit some confluence of unfortunate events to kick it off ... and so fixing it is going to require untangling what exactly that is, how it confuses the backend system, and then figuring out how to both untangle it now, and fix it so broken player accounts regain access to the chests. Plus, since nobody loses out on anything but time (and, well, a bit of profit, but such is life) on this -- key collection still works, and they can be spent when restored -- it has less urgency than if, say, it was a daily event that couldn't be caught up on.

Slippy, thanks for trying to help here. Does the router matter? Why not allow players to restart from a checkpoint? I'm poking at a theory that there may be a common element of which router brand is in use, and it may be the root cause of the disconnections by inappropriately dropping the NAT mapping for the TCP connection after 20 seconds of no traffic. Which, for almost every other use of the protocol, would not happen. (Though it would show up as drops for SSH connections, and given the 15-second protocol level pings to avoid this advised in places, would not be the first time this has popped up...) To the later: I'm all in favor of ANet allowing people to restart from a checkpoint. Sounds great. It wouldn't solve the problem, in the sense that if you are disconnecting now, you would disconnect again and again until you got lucky (or whatever it is). All it would do is reduce the amount of time you had to spend getting from the checkpoint to the disconnection point each time. So, regardless of checkpoints, I think the underlying issue should be investigated and solved. I'm all for things that improve the experience, and don't think these are in any way mutually exclusive options.

I'm curious, what brand and model of router do you have?

I'd assume that this is not nearly so simple as it sounds, because otherwise it kind of is the obvious thing to do. It might not be the easiest thing to do, but at least it's something to do. A feature like that would likely consume significant development resource meaning they'd have to push other engine changes (not content changes, that's most likely a different team). Would have knock on effects on other development pushing back other bug fixes or even the next expansion by however long it takes. I doubt there's much of GW1 in GW2, so it's not like restoring some old functionality, it's probably a whole new feature that may not sit well with the server infrastructure. The change also brings high risk of unexpected consequences I'd imagine, necessitating lots of testing. It's the kind of thing they should have built in to the game day 1.I mean, this is the obvious concern. The underlying technology clearly exists, but how easy it is to make that work with these story instances, I have literally no idea. My speculation is that it must have some hidden challenge, because otherwise it would be the obvious change, and the ANet team are no fools: if they could make this all go away quickly, why wouldn't they... We do know that only a small number of players are hit by this, but that it is very persistent for them. We know that an ANet developer is looking at this, though I can't imagine "full time" describes it. (I know that, given what they said about it, I usually instrumented things and waited for data collection to figure out what the common pattern was for these people, which isn't a full time thing, for situations where I dealt with similar issues in my own work.) Yah. So far we know that there is a common pattern in that the traffic on the TCP port used between client and server goes silent for 20-30 seconds, then there is no response from the server. I'm curious, for those hit by this, what brand and model of router y'all using, to see if there is a pattern there.

I'd assume it would only work if it was specifically within GW2, not some additional third party software. Hrm. What brand and model router do you use?

I'd assume that this is not nearly so simple as it sounds, because otherwise it kind of is the obvious thing to do.