Steam Thuesday maintenance

[xAv]

It would be nice if the steam thuesday maintenance (operate at 00h00 each thuesday Paris time), to don't kick player on desktop when playing a ranked game,
use loadbalancer please, and optimize your maintenace process

[❤ Sly Succubus ❤]

Автор сообщения: xAv

It would be nice if the steam thuesday maintenance (operate at 00h00 each thuesday Paris time), to don't kick player on desktop when playing a ranked game,
use loadbalancer please, and optimize your maintenace process

If you know maintence is coming up, why even join into a ranked game at all right before it?

[cSg|mc-Hotsauce]

Автор сообщения: xAv

Steam Thuesday maintenance

It would be nice if the steam thuesday maintenance (operate at 00h00 each thuesday Paris time), to don't kick player on desktop when playing a ranked game,
use loadbalancer please, and optimize your maintenace process

Maintenance usually occurs 4-6pm Pacific Time, near the end of the work day for Valve in Washington state.

This is done intentionally.

https://help.steampowered.com/en/faqs/view/7016-8AFD-19F7-F397

[[?]legit]

Автор сообщения: xAv

It would be nice if the steam thuesday maintenance (operate at 00h00 each thuesday Paris time), to don't kick player on desktop when playing a ranked game,
use loadbalancer please, and optimize your maintenace process

That would be great.

[Boblin the Goblin]

Автор сообщения: ?legit | Controller <3

Автор сообщения: xAv

It would be nice if the steam thuesday maintenance (operate at 00h00 each thuesday Paris time), to don't kick player on desktop when playing a ranked game,
use loadbalancer please, and optimize your maintenace process

That would be great.

They literally perform maintenance at the lowest player count.

[rawWwRrr]

Автор сообщения: xAv

It would be nice if the steam thuesday maintenance (operate at 00h00 each thuesday Paris time), to don't kick player on desktop when playing a ranked game,
use loadbalancer please, and optimize your maintenace process

You do realize that people around the world aren't living in the same time on the clock as you are, right? Even flat-earthers recognize that timezones exist, but on a ball-earth it's even more pertinent to understand that midnight where you live is a different time in 95% of the rest of the world. So why should Valve change the time of their Tuesday maintenance after nearly 20 years just to serve *your* convenience?

[xAv]

Автор сообщения: rawWwRrr

Автор сообщения: xAv

It would be nice if the steam thuesday maintenance (operate at 00h00 each thuesday Paris time), to don't kick player on desktop when playing a ranked game,
use loadbalancer please, and optimize your maintenace process

You do realize that people around the world aren't living in the same time on the clock as you are, right? Even flat-earthers recognize that timezones exist, but on a ball-earth it's even more pertinent to understand that midnight where you live is a different time in 95% of the rest of the world. So why should Valve change the time of their Tuesday maintenance after nearly 20 years just to serve *your* convenience?

don't be bitter ...

i don't want Valve change time of his thuesday maintenace, i just want valve use modern process, to execute them, shutdown all server at the same time for operate maintenance, is a non sense, we are in 2023, they must change their prehistoric IT Teams, you imagine google shutdown all their servers around the world at the same time ???

you may not have any knowledge of loadbancers, other large companies know how to perfectly manage maintenance deployments without service interruption

[Mad Scientist]

Автор сообщения: xAv

i don't want Valve change time of his thuesday maintenace, i just want valve use modern process, to execute them, shutdown all server at the same time for operate maintenance, is a non sense, we are in 2023, they must change their prehistoric IT Teams, you imagine google shutdown all their servers around the world at the same time ???

Valve is their own entity, and like most entities they figure out what works best for them including doing live maintenance while people are present at the company. If you're used to it happening on tuesdays, stop playing games reliant on the systems rather than a Developers own systems around the time of maintenance. Their practices work best for them. Live maintenance for them disrupts as little as possible, which also affects a small window of time typically.

Doing what others do is not the best practice, when a company knows what works best for themselves, it's definitely not "prehistoric" and anyone competent in this field would recognize that.

[xAv]

Автор сообщения: Mad Scientist

Автор сообщения: xAv

i don't want Valve change time of his thuesday maintenace, i just want valve use modern process, to execute them, shutdown all server at the same time for operate maintenance, is a non sense, we are in 2023, they must change their prehistoric IT Teams, you imagine google shutdown all their servers around the world at the same time ???

Valve is their own entity, and like most entities they figure out what works best for them including doing live maintenance while people are present at the company. If you're used to it happening on tuesdays, stop playing games reliant on the systems rather than a Developers own systems around the time of maintenance. Their practices work best for them. Live maintenance for them disrupts as little as possible, which also affects a small window of time typically.

Doing what others do is not the best practice, when a company knows what works best for themselves, it's definitely not "prehistoric" and anyone competent in this field would recognize that.

Yes you are probably right, an twenty years old method is certainly better than recent one ,
it seems to not possible to exchange about that here, really a shame to have such restricted exchanges, without wanting to progress for the comfort of all

Have a nice day

[Mad Scientist]

Автор сообщения: xAv

Автор сообщения: Mad Scientist

Valve is their own entity, and like most entities they figure out what works best for them including doing live maintenance while people are present at the company. If you're used to it happening on tuesdays, stop playing games reliant on the systems rather than a Developers own systems around the time of maintenance. Their practices work best for them. Live maintenance for them disrupts as little as possible, which also affects a small window of time typically.

Doing what others do is not the best practice, when a company knows what works best for themselves, it's definitely not "prehistoric" and anyone competent in this field would recognize that.

Yes you are probably right, an twenty years old method is certainly better than recent one ,
it seems to not possible to exchange about that here, really a shame to have such restricted exchanges, without wanting to progress for the comfort of all

Have a nice day

Yet the disruptions are far less than they used to be, so it has improved.

Something refined over time is not the same as the original, it is merely an identical functional route taken to update & resolve issues. Since tis refined, it's clearly still efficient with minimal downtime overall.

[[?]legit]

Автор сообщения: xAv

Автор сообщения: Mad Scientist

Valve is their own entity, and like most entities they figure out what works best for them including doing live maintenance while people are present at the company. If you're used to it happening on tuesdays, stop playing games reliant on the systems rather than a Developers own systems around the time of maintenance. Their practices work best for them. Live maintenance for them disrupts as little as possible, which also affects a small window of time typically.

Doing what others do is not the best practice, when a company knows what works best for themselves, it's definitely not "prehistoric" and anyone competent in this field would recognize that.

Yes you are probably right, an twenty years old method is certainly better than recent one ,
it seems to not possible to exchange about that here, really a shame to have such restricted exchanges, without wanting to progress for the comfort of all

Have a nice day

Well said. To be honest, I can neither agree or disagree with him. What is he even saying?

I will support your suggestion though, not getting kicked during an online session every week would be great.

Автор сообщения: Mad Scientist

Yet the disruptions are far less than they used to be, so it has improved.

That doesn't change anything about the problem.

Автор сообщения: Mad Scientist

Something refined over time is not the same as the original, it is merely an identical functional route taken to update & resolve issues. Since tis refined, it's clearly still efficient with minimal downtime overall.

Okay.. xD

[RiO]

Автор сообщения: Mad Scientist

Автор сообщения: xAv

i don't want Valve change time of his thuesday maintenace, i just want valve use modern process, to execute them, shutdown all server at the same time for operate maintenance, is a non sense, we are in 2023, they must change their prehistoric IT Teams, you imagine google shutdown all their servers around the world at the same time ???

Valve is their own entity, and like most entities they figure out what works best for them including doing live maintenance while people are present at the company.

What OP is asking for is not to shift the time of the maintenance.
What they're asking is to use modern practices for updating public-facing services that avoid service disruption.

The common -- yes common; Valve is the archaic outlier here -- way to handle this is:

1. take half of your server cluster out of your ingress load balancer;
   
2. wait for traffic to those servers to stop (matter of seconds; just finishing open requests);
   
3. update those machines and ensure everything works with self-tests;
   
4. switch over part of the real-world traffic to the updated part of the cluster;
   
5. use that traffic to verify that everything also holds under real-world conditions;
   
6. shift everyone over to the updated half;
   
7. let traffic to the other half die out;
   
8. update the other half of the cluster;
   
9. same process; verify everything; shift part of the traffic back to it; etc.
   
10. then finally reintroduce everything back into the ingress load balancer.

If you do it right; you can run regular scheduled maintenance without your customer base noticing any down time whatsoever; or with minimal impact where issues do occur.

Doesn't have to be a half-and-half system either. You can employ it with any working size that works well for you.

[Mad Scientist]

Автор сообщения: RiO

(snip)

Without noticing any downtime or affected systems is incredibly difficult even for the most competent programmers. Doing what works best for ones own company is entirely acceptable. Either way something will be affected, every company I've worked for or dealt with has always had something display towards the most active employee or user, else something fully reliant on not being disrupted even in the slightest. There's too many variables, that no matter what they do someone will be affected.

Others use long complete downtime (ie nothing live at all other than the basic website functions) which is immensely more disruptive.

It's virtually impossible for someone not to notice due to the active userbase, but how extremely little of a window it usually affects any one thing speaks highly of overall competence.

[Nx Machina]

Автор сообщения: RiO

The common -- yes common; Valve is the archaic outlier here -- way to handle this is:

Valve's platform, Valve's choice.

Secondly Blizzard takes their servers offline when it is not convenient for me, because again like Valve, it is their platform and their choice, after all like Valve they have ALL the data to know what is best for them.

And finally Ubisoft an EU company takes their servers offline when it is not convenient for me, when i want to play The Division or The Division 2.

The outcome and solution in all three scenario's is to adapt.

[RiO]

Автор сообщения: Nx Machina

And finally Ubisoft an EU company takes their servers offline when it is not convenient for me, when i want to play The Division or The Division 2.

Ubisoft's servers go down on the regular, maintenance or not.
They're up when you're not looking. They go down as soon as you want to play anything that needs them.

They should ask Gene Roddenberry's estate for the schematics to the Heisenberg Compensator. I mean; he was / they were friendly enough to hand Apple the ones for the tricorder to give them a leg up with the iPhone.

Fail to see the relevance anything EU has here.
Unless it's meant to try and bait something, or what?

[[?]legit]

Автор сообщения: Nx Machina

The outcome and solution in all three scenario's is to adapt.

Yes - Valve should adapt to modern practices, instead of users adapting to medieval practices.