Xbox One Xbox One Cloud Updates May Force Servers to Reboot Mid-Game

KN1GHTMARE

KN1GHTMARE

Former Moderator
Joined
Aug 31, 2013
Messages
1,933
Karma
32
From
NJ
Website
socialclub.rockstargames.com
Microsoft has announced the Xbox One will require cloud updates on a regular basis, which could lead to the system rebooting without warning. That's according to Xbox Live lead program manager John Bruno, who made the revelation during a presentation at GDC NEXT, yesterday (via IGameReponsibly). Bruno explained that while work has been done to try and prevent the system rebooting in the middle of a game, Microsoft hasn't managed to eliminate the issue entirely.

"Once in a while, rather frequently actually, the host OS will require an update, meaning the physical machine is going to get rebooted whether your code is running or not," he explained. "That’s a problematic thing for a game, and oftentimes it's in the middle of a multiplayer session. We’ve worked very hard to overcome that, but that’s not to say it’s going to be a reality in every case."

It's possible, then, that the cloud may have a weekly maintenance period to try and avoid issues, scheduled during a convenient time for most gamers, such as in the early hours of Wednesday morning like many MMOs. We've reached out to Microsoft to see if there are any such plans in the pipeline and ask what happens if an emergency maintenance is needed, and will update the story as soon as we hear back
Click to expand...

Well this is definitely troubling information. I knew the first few days would be rough, but this is upsetting and seems to be an ongoing problem with no solution yet in sight. It will be very frustrating to be booted out of our multiplayer sessions for an update and have no choice over the issue. Considering multiple reports previously had said the OS and XBOX LIVE were in shambles, lets hope there's not more bad news headed our way.

Update: Multiple sources are now saying the console will not reboot, but rather the Azure cloud servers. This will only affect applications using the cloud servers at the moment, meaning like Titanfall. The physical machine referred to is the physical server not the console

via IGN
 
I'm sure there will be something MS can do to fix/halt that while the game is in session, then simply move it to another server.
I doubt with all sincerity, that the entire cloud would reboot at once, you know?
 
Ok, I'll say a bit about this, because I'm an engineer who uses Azure every day. In fact, I reboot instances just about every day as well.

First off, the articles popping up about this are completely misleading, as mentioned above. It's the instances that get rebooted, not the console.

When you think of Azure Instances, don't think of them as servers. Instead, they are more like CPU/RAM capacity, where each instance increases capacity. An Azure server is actually called a RIG. Each RIG can contain many Instances, those instances are grouped together by service/software (example: a group for Titanfall and another group for Forza 5 might all be a part of the same RIG).

So when you remove (reboot) an instance you aren't losing data, but instead are just losing overall capacity.

To further explain: Lets say you have 5 instances grouped together running Forza 5, the group is able to handle 100,000 connections (each instance allows the instance group to handle 20,000 connections). But lets say only 50,000 people are actually playing/connected to that group of instances.

Now let's say you reboot 1 of those 5 instances, you don't lose connection, but your overall capacity drops down to 80,000 connections until it's fully rebooted. However, since only 50,000 people were connected nobody is disconnected while the instance reboots.

Now what I think John Bruno is saying is that the RIG itself gets rebooted (causing all attached instances to drop until the RIG is rebooted). This is the same situation we have today, where if a server is rebooted it will drop connections, which is why server reboots have always been done at off-peak hours and each server is first drained of connections first (technically they just refuse new connections by taking the server out of rotation and once the 'total connected' drops below a pre-determined threshold - often zero, the server is then rebooted). However, from my experience working with Azure, the RIGs almost never get rebooted...and even the RIGs are not 'servers' in the traditional sense.

Overall I have no idea why John Bruno bothered to speak on the issue. He's a PM, not an engineer, so he isn't someone who actually works with Azure patching/reboots. That's something that I do, however.
 
Who wants to bet that the other camp is going to RUN with this even though its misleading lol
 
I'm holding off all complaints and worries until I have system in my hands. All these stories are about an unfinished non-updated system :).
 
KCJRMC said:
I'm holding off all complaints and worries until I have system in my hands. All these stories are about an unfinished non-updated system :).
Click to expand...


I am the same. Seems the closer we get to launch, the more negative some people are getting and the more negative type things pop up.
 
I was worried when I first read this article, but I figured MS had a contingency plan in place when this maintenance occurred.

J
 
we shall see.. i get dropped signal/internet from time to time so it wont feel unusual to me lol.
 
Landshark said:
First off, the articles popping up about this are completely misleading, as mentioned above. It's the instances that get rebooted, not the console.
Click to expand...

Misleading articles with sensationalistic headlines? In my games journalism? That's a shocker.
 
Without warning? I know they're working hard to fix it, but until they get that issue resolved there should at least be a little warning. Something like, "oh, by the way, this server is rebooting for an update in about 5 minutes, save your stuff".
 
This sounds a bit troubling but I also feel like there's not a whole lot of information so I'm just going to and see how it works.
I definitely don't want to be playing a multiplayer game only to be suddenly disconnected and lose all my progress until then.
Let's just wait and see how for ourselves how the console works and then we'll talk.
 
It just shows how people are looking to go after MS any chance they get and will run with misinformation and fear mongering. It makes you wonder though why MS doesn't tell employees to watch what they say. Landshark knows a lot more about it and explained it well. All I know is working in IT for going on 20 years that something simple like this will be handled. You can do rolling reboots on off peak hours, scheduled maintenance windows, etc. 24/7/365 availability is something that is very easy to maintain.
 
In my eyes those are just rumors that were published to make people choose another console instead of one. And as the info source is not some trustworthy rep, I think this is either fake or it is a TEMPORARY problem that ms will fix before releasing one, but someone wanted us(potential buyers) to be afraid of that bug. There is no way MS would release something like that.
 
Kerosene31 said:
It just shows how people are looking to go after MS any chance they get and will run with misinformation and fear mongering. It makes you wonder though why MS doesn't tell employees to watch what they say. Landshark knows a lot more about it and explained it well. All I know is working in IT for going on 20 years that something simple like this will be handled. You can do rolling reboots on off peak hours, scheduled maintenance windows, etc. 24/7/365 availability is something that is very easy to maintain.
Click to expand...

Honestly sometimes I wish there was a way for people who have no IT or Operations experience to get face-to-face time with what really goes on in a global service infrastructure setup. Having done this type of work for the past 10+ years (at a corporate level - 6 of which have been at Microsoft), it can be both frustrating and entertaining to read people's posts where they 'think' they know what goes on, but really have no clue. In a lot of cases it ends up causing blame being placed entirely wrong.
 
Landshark said:
Ok, I'll say a bit about this, because I'm an engineer who uses Azure every day. In fact, I reboot instances just about every day as well.

First off, the articles popping up about this are completely misleading, as mentioned above. It's the instances that get rebooted, not the console.

When you think of Azure Instances, don't think of them as servers. Instead, they are more like CPU/RAM capacity, where each instance increases capacity. An Azure server is actually called a RIG. Each RIG can contain many Instances, those instances are grouped together by service/software (example: a group for Titanfall and another group for Forza 5 might all be a part of the same RIG).

So when you remove (reboot) an instance you aren't losing data, but instead are just losing overall capacity.

To further explain: Lets say you have 5 instances grouped together running Forza 5, the group is able to handle 100,000 connections (each instance allows the instance group to handle 20,000 connections). But lets say only 50,000 people are actually playing/connected to that group of instances.

Now let's say you reboot 1 of those 5 instances, you don't lose connection, but your overall capacity drops down to 80,000 connections until it's fully rebooted. However, since only 50,000 people were connected nobody is disconnected while the instance reboots.

Now what I think John Bruno is saying is that the RIG itself gets rebooted (causing all attached instances to drop until the RIG is rebooted). This is the same situation we have today, where if a server is rebooted it will drop connections, which is why server reboots have always been done at off-peak hours and each server is first drained of connections first (technically they just refuse new connections by taking the server out of rotation and once the 'total connected' drops below a pre-determined threshold - often zero, the server is then rebooted). However, from my experience working with Azure, the RIGs almost never get rebooted...and even the RIGs are not 'servers' in the traditional sense.

Overall I have no idea why John Bruno bothered to speak on the issue. He's a PM, not an engineer, so he isn't someone who actually works with Azure patching/reboots. That's something that I do, however.
Click to expand...

I was about to say this.

Ok...no I wasn't because I have no idea what you mean but I trust you know what you are saying hehe.
 
Landshark said:
Honestly sometimes I wish there was a way for people who have no IT or Operations experience to get face-to-face time with what really goes on in a global service infrastructure setup. Having done this type of work for the past 10+ years (at a corporate level - 6 of which have been at Microsoft), it can be both frustrating and entertaining to read people's posts where they 'think' they know what goes on, but really have no clue. In a lot of cases it ends up causing blame being placed entirely wrong.
Click to expand...

If people really knew that what we did was not that hard, we wouldn't get paid nearly as well! :cool:
 
MS has already clarified this (not sure why they let their people put stuff like this out there to begin with)

http://www.techradar.com/news/gamin...g-gameplay-after-all-1198532?src=rss&attr=all

Microsoft has sent TechRadar a statement to clarify that no, the Xbox One will probably not reset while users are in the middle of a multiplayer match.

Last week Microsoft's Xbox Live Lead Program Manager John Bruno made some comments during a GDC Next talk that many interpreted to mean that server-side updates could cause the Xbox One to do just that.

But Bruno's words were taken out of context, a Microsoft spokesperson clarified in a new statement to TechRadar.

That's good news for future Xbox One gamers who were worried about getting shut down in the middle of a Call of Duty match.
Click to expand...
 
You must log in or register to reply here.

Latest posts

Latest threads

Back
Top