Distant Worlds II Event Servers offline after the first jump of Distant Worlds 2. Come on Frontier!
- Thread starter Maro-Val
- Start date
there are 3 launches taking place, so it's not 10K people, this launch was for EUR people, US is tonight at 8EST, and AUS a few hours after that.
I think while Frontier had ample warning. I don't believe anyone of us could have predicted that there would be 11,000+CMDRs who signed up for the expedition.
I guessed this would happen. There were issues all weekend.
I can't wait for the US jump tonight.
I'm surprised it wasn't worse! The game didn't crash while I was in the instance, and for me, the jump worked. It wasn't until I started to plot my route in the other system the game hung-up on me. But the game didn't crash, which is a first for me. All the other mass jumps I've been part of crashed the game or froze it up. This time I was just kicked out to the menu. A slight improvement, I think.
I decided to fly back to Panini, Pallanin, Palpaltini, or whatever it's called, and join the next jump event (unless our dinner guests have arrived by then).
I decided to fly back to Panini, Pallanin, Palpaltini, or whatever it's called, and join the next jump event (unless our dinner guests have arrived by then).
Last edited:
Reasonably well, although I am more of an Azure guy.
However, whilst I appreciate your comments, it's not me that made claims about using AWS to scale like this, It's Amazon and FDev themselves. Amazon even use Frontier as a case study;
https://aws.amazon.com/solutions/case-studies/frontier-games/ and they say "By using AWS, Frontier Games can scale compute resources easily to handle large spikes in user traffic"
Here's the presentation; https://www.slideshare.net/AmazonWe...realtime-commodities-market-aws-reinvent-2015 - Have a look at Slide 18. They make bold claims about the autoscaling infrastructure - using elasticache and auto scaling groups in front of the RDS you mention. Notably, they still refer to this environment as "massively multiplayer".
You mentioned on another post about the rate of change being the thing - and that's why I am annoyed. I have been involved in the architecture of two major systems that are massively scalable - one which had to cope with a massive influx (150k+ users) all at pretty much exactly 8.30am, and one which responded to the "stick the kettle on" ad break in major soaps.
In both cases, the admin team pre-warmed FE instances to cope with demand rather than letting the auto-scaling work completely on it's own. This event was well communicated ahead of time. FDev knew pretty much exactly within a few seconds when the spike was coming.
This should not have happened - and given the coverage of the event in the gaming press (hell, there has even been coverage in the non-gaming press), this absolutely should not have happened.
Seeing as things seem to have settled down I think it was a success.I was expecting a complete meltdown of servers and the community which turned out to be a server crash some brief moaning and then business as usual.
Well done all o7
Well done all o7
I think that, in honour of this occasion, Frontier should add some of those Q or P type (or other) Anomalies to float around the DW 3302 tourist beacon. Because that many thousands of hyperdrives all going off in the same place at the same time surely must have done something weird with space-time - maybe unleashed some anomalous life forms into the area.
Nice little marker of not only where DW 3302 started from, but also where DW2 broke reality.
Nice little marker of not only where DW 3302 started from, but also where DW2 broke reality.
So you usually let a little pee on on Sunday nights?
Ok I blame my instance[video=youtube_share;D-tTyGhy7gc]https://youtu.be/D-tTyGhy7gc[/video]
...and they were never seen again.
Last edited:
AFAIK, US launch went off with little drama.
If that is in fact the case, well done FD.
Now keep it all together for the Oceania launch, because that's where I'm launching.
If that is in fact the case, well done FD.
Now keep it all together for the Oceania launch, because that's where I'm launching.
Agreed there were between 5K and 10K concurrent connections at the time of the jump (if we look at DW2 registration numbers, optimistically) that's absolutely nothing compared to some production servers running for 100K+ users connecting at the same instant. AWS by itself can handle that perfectly. I suspect the network code of Elite to be quite unhealthy.. (also, distributed write caches are a thing..)
No. Instances were the usual ~20 max (around there).
Last edited: