DPG - Please Explain the Technical Side of Aradune's Problems

IIRC they posted that it wasn't a physical migration. My guess at the time is they were using an Oracle DB and with the changes Oracle did with licensing you are seeing a lot of people move to some type of Open DB like postgres.

Got a quote from that? I'm hearing there was a physical migration from inhouse servers to a server farm, they went cheap AF on their selection, had some personnel losses that are preventing them from restructuring or scaling in any way and are currently incapable of providing service for very active servers.

But that's second hand info.

My theory is that they tuned the Aradune server resources for the projected player population they will see after the initial launch boom. This post-launch population is significantly lower than classic and seems to be relatively consistent from Kunark to PoP era. The advantage of this is that they're bad at managing infrastructure so the idea of a potentially catastrophic upsizing then downsizing of the server is avoided but the drawback is that the initial launch population experiences queues and long zone times. I think they did this with Mangler which is why we saw some hiccups during launch and instability throughout the life of the server and I suspect they downscaled things further for Aradune.

Trace your network and all the IPs it’s hitting are locating in San Diego and owned by DBG. Could be routing all traffic from some server farm back to them...or they are still managing their own servers.

Now what’s interesting is each zone seems to have its own IP along with chat having a separate IP. That to me seems to be the very old way of having a dedicated section of a machine per zone as opposed to the just in time method they later came up with. Both are talked about https://www.google.com/amp/s/spectr...tronics/gaming/engineering-everquest.amp.html

“ A few years ago, every time a player in a particular world entered a specific virtual dungeon, chances were the same central processing unit, dedicated to that place alone, served up the data seen on the display. Each CPU of each server in a world cluster was assigned a geographical task. But that way of dividing up computing resources turned out to be too inefficient to handle the new content the game’s designers were pumping out. So a few years ago, Sony shifted to a new way of managing the system: just-in-time computing.”

What do you imagine their database to be? Not that long ago it was flat files. I have no idea what it is now, but I wouldn't assume it's SQL Server, Oracle, etc.

Not that I can find off searching for 5 mins and not invested enough to search more. Just going of memory here of what I recall reading back when it was happening.

They talked about it at GDC a few years ago. How they hired an ex-NASA engineer to convert a lot of content from flat files to a DB. I can't recall if they said what DB they used, nor did they say exactly how much or what content they migrated. Although in another thread about server merges it was confirmed your char info is now stored in a DB instead of a flat file, so that is a huge change.

There is no recording of the panel they did talking about this either so can't go back and look.

nm, found it.

It is no exaggeration. I got group invite in Unrest yesterday, zoned to their pick... hit the ever-zoning screen... waited like 2 minutes real time... went and fixed a chicken salad sandwich... ate it while watching the everzone screen... 30 min later CTD.

That quote reads to me as they're not physically moving servers to a new location but just pushing the data to another location.

That's how I read it too. They're not moving their inhouse server farm to another building, they're spinning up a third-party server farm by transferring the data to it.

Tangential derail, I used to work with an engineer who had also worked at NASA. His time there was only around 5 years and only part of a solid career. He was annoyed that he was constantly introduced or referred to as having worked for NASA. After I learned that it was hilarious every time someone brought it up. We'd meet some dudes, a boss-type would say, "Hi guys, here's Fippy Darkpaw. He's an ex-NASA engineer." and I'd glance at him to see the restrained annoyance.

Good info, thanks. I vaguely recall that they moved to a real DB at some point.

We hired an ex-SpaceX guy a decade or so ago. Different subject, but I was surprised to hear from him that their software processes back then were pretty bad. I don't recall the details but at the time I remember being shocked. A lot has changed there in 10 years probably.

Yeah see I take it differently. Anytime you are moving to different hardware it has usually always been refereed to as a physical move. Migrating data from one DC to a different one would be a physical move in my books. But who knows what was actually done here. Could have been a combination of migrating to an opensource DB (makes sense to reduce costs) as well as move to a cheap IaaS vendor.

lots of people, really cheap server. nuff said.

I agree that there should be an Aradune #2, but " 'just' do a second Aradune" implies that adding another server is easy within their current resources (human, infrastructure, financial). I wouldn't be surprised if they're planning on doing just that, but I'm also sure they're wary of investing in another server only for demand to die off in a couple more weeks (from the max level and leave crowd) and then they're stuck with the need to merge.

Humans are hard to predict and attempting to do so is a gamble.

While I'd be ok starting over, I want the bags I purchased for Aradune 1.0 to transfer, or be claimable again, to Aradune 2.0.