LOL snorted a bit on that one
1 Like
You, too, were there when the dark magics were written!
1 Like
Why use Oracle 
They do its PROD lol not UAT or DEV they just straight push that shi
1 Like
We in 2024 now, TWW been out for 5 weeks + early access. Eventually the copium will need to die
Is this ChatGPT?
ie. this is completely wrong, they don’t push update across all regions at the same time
WoW operates across different regions, including North America, Europe, and Asia, each with its own infrastructure. During maintenance, these servers need to be synchronized to ensure that the changes applied are consistent across all regions.
2 Likes
Afraid we can’t do that. Explaining maintenance will force us to extend the maintenance window by another three hours.
You’re assuming they know.
Not fair! All my chars are already loaded up on champion and hero sacbroods. maybe if they passing out mythics…
A number of reasons, but I think the simplest way to put it is this:
WoW’s a live service that runs 24/7 to give a near constant and persistent experience, so that means the hardware is running that entire time. When you have something that operates constantly, you need time to stop it and make sure it can continue running that effectively in order to catch problems before they occur. Sometimes there need to be changes in hardware, software bugs need to be squashed, the list can go on really.
When you have a window of time extended like this, it probably means they’re encountering a problem that’s been causing consistent issues. It could be a number of things, ranging from more complicated hardware issues, a logistical one, or hell, even a laborforce issue.
2 Likes
Sometimes we have small updates, hotfixes, that don’t require bringing down the game servers. But for the larger updates, here’s some possibilities.
–
Imagine you have a candy business and you need to make some changes to the physical store.
–
- (Database Migrations) Moving candy jars: If we need to move a jar to a new shelf or get rid of some candies that don’t fit anymore, we have to close the store for a bit to make sure we don’t spill anything.
- (Monolithic Architecture) One big candy counter: If the whole store is built around one big candy counter, and we want to change even one small thing, we have to close the whole store while we fix it.
- (Deep System Changes) Building new shelves: If we need to build new shelves or fix the floor, we can’t have people walking around. We need to close so we don’t make a mess or trip anyone.
- (Data Consistency) Counting candy carefully: If we need to count all the candies and make sure everything is in the right place, we have to close the store for a bit so nobody buys or moves anything while we’re counting. That way, everything stays correct.
–
Alternatively, there are methods to minimizing downtime, but my guess is that the cost (money, human resources, etc.) are deemed too great to Blizzard.
–
Imagine instead of closing the candy store, we build a new, identical candy store right next door:
- While the old store is still open and selling candy, we quietly work on the new store. We move candy, set up new shelves, and make everything look great.
- Once the new store is ready, we close the old one for just a minute and say, “Hey, everyone! Come to the shiny new candy store!” All the customers walk right in, and they can keep buying candy without waiting.
This way, we don’t have to shut down the candy business for a long time. The switch happens quickly, and everyone’s happy with the new changes!
This requires a lot of the infrastructure to be duplicated during this process, which is probably why it doesn’t happen here. It takes a lot of work to set something like this up at this scale.
–
My opinion. None of this might be true.
1 Like
They did not appease the machine spirit
1 Like
do you think they’ve tried turning it off and on again?
3 Likes
Somebody deployed a helm chart with the wrong value and they can’t figure out what it is.
They are calling Blizzard IT to fix the actual problems.
2 Likes
Thanks ChatGPT
1 Like
the servers get dirty over time, they’re just using some tools to clean them up
I believe it’s more like: git reset --hard
git rebase
1 Like
I’m a retired 40 year developer. Its how I experienced it.
Large and very old projects are too big to re-write and too old to know properly so new guy after new guy puts, “BANDAIDS” in to get the job done on ask after ask. Given a few generations of new guys and you have what I call “an app written by Johnson & Johnson” in other words, more Band-Aids than app. Then one day someone is asked to make a simple fix and happens to pull an important band-aid off and it break all kinds of things making a massive mess and no way to roll it back so they scramble to get it back up by throwing in dozens more band-aids (Software industries version of duck tape).
Keep in mind, your not allowed to say you see a potential problem because that would make you negative. You have to simply watch it happen and act surprised.
That’s how it was for me.
6 Likes
they submitted a ticket
2 Likes