Cute, but none of this requires hours of downtime and most of it can and should be done during off-hours (e.g., 3:00 am server time).
how do you know those are off-hours? do you have numbers to back that up?
wut thats insane! free character transfers?
Obviously whenever Zups is not playing.
1 Like
You don’t need statistics if you’re patching.
Source: Statistics
1 Like
No joke, back when I did more field IT work for a company I used to work for, I had SO MANY TICKETS that I’d resolve just by doing that due to some coroutine or other process getting stuck, so rebooting it just resolved it.
Whenever it came to servers though, hoo boy. Any time we would get a new one, we would have to set it up days in advance so it could start pulling the current build from our headquarters. Worst part about that? Sometimes we had VERY limited connection access (we’d have to use 4G sometimes), and some of the server builds were so big that they’d take days to finish updating entirely.
The servers have underlying hardware and operating systems.
Those systems need to get updated with security patches, and that typically means at least a reboot.
Server performance is (hopefully) monitored, so there may be some things that require downtime to upkeep. You can often hotswap hard disks in raid arrays, but it’s not particularly advisable. RAM that’s aging or malfunctioning, you’re going to need to power off to replace.
The servers probably also use a hypervisor, so if that needs updating, that’s going to require downtime.
In a ransomware world, you are going to want a mix of live and offline backup data, so it’s probably that live backups are rolled over to offline backup during these windows also.
I think you’re trying way too hard sir.
I doubt it’s an OS issue. More than likely the codebase encountered some type of error. I blame these transactional systems they’ve implemented (Trading Post, in game shop, etc.) after/outside the game’s codebase that’s trying to interact.
They just need to download more ram, duh
Why would they waste the effort explaining it?
Either way this forum will complain and it’s eating their time on top of it - it’s a no win proposition for them to explain what they are doing.
No to mention who would really care if they says “well some of the Gbics on a few switch stacks need replacing then there’s firmware upgrades on select servers , so hopefully after that the database dies not take too long…”
People here don’t care they just wanna play - they’d be like ok fine how much longer?
Test labs are simply not at the scale of production.
Further, sometimes the hardware itself needs to be changed. (One day I put new RAM in my computer. Everything ran fine until a memory test failed.)
Can attest. My test labs may hit a couple hundred users and that’s with them somewhat “real users.” You can’t account for production users and data overwhelming it. This is a “getting it mostly right” type of deal and you patch the rest.
Add another hour because someone not even on the approval request is now blocking due to semantics
1 Like
LOL… 14.4K modems my man!! We only had 0G in the 90’s. LOL
You win the interwebz today.
1 Like
what makes you think there’s a codebase issue?
you should be doing regular maintenace monthly if not weekly for patches. guaranteed.
patching game-related bugs doesn’t require fixed schedules like this and can generally be done on live systems.
The number of times I’ve pointed out how an approach is going to lead to problems is so frustrating. Most of the times, they can’t see it. Then N months later disaster strikes, you tell them “this is what I told you would happen”, but they don’t remember because it’s like you were speaking another language. You can’t retain the meaning of things you don’t understand.
God it FEELS like we’re downloading it on like AOL speeds sometimes XD
ZERO idea why the contractor doesn’t just have the network guys come in first. It’s one of the weirdest things when we get put on these particular projects.
The only real reason they’d need to actually restart the servers - is a memory leak problem.
Otherwise, restarting a machine is riskier than just leaving it run.
And - there’s a super high likelihood it’s server/database folks needing this time, and not actual software engineers.