mips_winnt wrote: »Actually that's not the BEST way to do it, the best way to do it is to have redundant systems where you do a stateful failover to your backup systems while you implement changes on your primaries. That way your customers aren't suffering an outage while you're implementing changes and if your changes blow up your production systems you have an easy rollback, this is how most enterprises operate their infrastructures with respect to customer facing systems.
Posivatedb14_ESO wrote: »I'm just curious on what other peoples opinions are on having a maintenance every week. We pay our monthly fee of 15$ and are not able to play 4-10 (Sometimes more or less) hours a week due to maintenance. I would think in 2014 we would be beyond having to bring a server down for hours at time, and at a time that conflicts with some peoples schedules every single week.
I know this topic has come up much in other MMO forums, but I was hoping ESO would be different. I think of this like having a car that you pay for every month that requires you to not use it for a day every week of its existence. What are your thoughts on this? I am also well aware this won't change anything, I'm just interested to see what other's think.
Kyotee0071 wrote: »Downtime is expected.
Downtime should Ideally be done in the hours of non optimal playtime for the masses.
Example for a NA server - 4 AM EST. Now I know there are folks that 4AM EST would be their "prime time", but for the majority of the people (opinion not fact) on the NA server that would be a time they wouldn't notice the downtime as much.
I never understood why companies do their maintenance during Mid day / Evening hour time frame of the server.
Again, I'm not complaining - just giving my opinion.
jwwoffordb14_ESO wrote: »The issue is not with scheduled maintenance, but the fact it is done during two WEEK DAYS during "business hours". They need to get with the program do that work late, late at night or early early in the am (like 4 am). I predict that new gamers are going to want an Amazon.com, or Gmail.com level of service for their $15 per month fee and I do not blame them.
Do the patching/reboot/maintenance in the middle of the night, on the weekends like the rest of IT! The techs can sleep in late after the work gets done.
That is not accurate in a virtualized environment, hardware capacity does not present nearly the constraint that it once did and replicating a complete (and up to date) production environment is not a steep technical or CapEx challenge.mips_winnt wrote: »Actually that's not the BEST way to do it, the best way to do it is to have redundant systems where you do a stateful failover to your backup systems while you implement changes on your primaries. That way your customers aren't suffering an outage while you're implementing changes and if your changes blow up your production systems you have an easy rollback, this is how most enterprises operate their infrastructures with respect to customer facing systems.
Thing is that those systems cost, and you have to make sure that the capacity of that system is the same as the normal park you have. A back-up system normally only have records of the data, it isn't however playable and doesn't have the server load capacity to handle the traffic,
It's apparent you're not familiar with how infrastructure virtualization works, I'm not talking about "fail safe" systems wrapped around redundant hardware, I'm talking about infrastructure that can be replicated in software and that is scalable on demand (see VMware, Citrix, OpenStack, etc.., etc,,,). Most business that don't have the heads completely back in the last decade use such technologies not only to optimize their hardware CapEx but also to minimize planned and unplanned downtime on their customer facing systems.it's a pure back-up system, nothing more. As I said, they would have to different server parks with the same capacity on the fail-safe system as on their normal system to not have a server-down situation, and that simply isn't viable in terms of cost.
They have - as all enterprises have - a network setup with back-up systems to take care of the files. This server however is build for that purpose and wouldn't have the capacity to handle the traffic of the players wanting to play.
That's not a problem, you patch/fix/restart your primaries while your backups service your clients, once your done with validation you fail back to your primaries (in the case of an MMO you can't do a stateful failback if you're requiring client side patches but it still means minimal downtime to your customers) ...... .Another problem is that some changes need a complete server restart for them to work. This is because the error isn't server based but client based and would require the client to be patched in order for the error to fix itself.
Only if your "technicians" don't know what they're doing.....So the best way for the technicians working on the servers are to take down the servers the time it takes to fix it.
MMO's only get away with it because the industry has managed to set customer expectations so low, in any other business the amount of downtime these jokers get away isn't tolerated (nor should it be), it isn't necessary with current off the shelf technology.Other MMOs do that all the time, and most - if not all - games with an online feature takes their servers down for maintenance now and then. Most have the schedule to do so with the least interruption, but others might not have that luxury.
That is not accurate in a virtualized environment, hardware capacity does not present nearly the constraint that it once did and replicating a complete (and up to date) production environment is not a steep technical or CapEx challenge.
It's apparent you're not familiar with how infrastructure virtualization works, I'm not talking about "fail safe" systems wrapped around redundant hardware, I'm talking about infrastructure that can be replicated in software and that is scalable on demand (see VMware, Citrix, OpenStack, etc.., etc,,,). Most business that don't have the heads completely back in the last decade use such technologies not only to optimize their hardware CapEx but also to minimize planned and unplanned downtime on their customer facing systems.
That's not a problem, you patch/fix/restart your primaries while your backups service your clients, once your done with validation you fail back to your primaries (in the case of an MMO you can't do a stateful failback if you're requiring client side patches but it still means minimal downtime to your customers) ...... .
MMO's only get away with it because the industry has managed to set customer expectations so low, in any other business the amount of downtime these jokers get away isn't tolerated (nor should it be), it isn't necessary with current off the shelf technology.
Yeah see it's obvious that not a lot of people who browse this forum are familiar with these concepts. To be honest, I wouldn't have very much idea about the state of virtualization had I not been attending a development conference this week. It's been quite an eye-opener.mips_winnt wrote: »That is not accurate in a virtualized environment, hardware capacity does not present nearly the constraint that it once did and replicating a complete (and up to date) production environment is not a steep technical or CapEx challenge.mips_winnt wrote: »Actually that's not the BEST way to do it, the best way to do it is to have redundant systems where you do a stateful failover to your backup systems while you implement changes on your primaries. That way your customers aren't suffering an outage while you're implementing changes and if your changes blow up your production systems you have an easy rollback, this is how most enterprises operate their infrastructures with respect to customer facing systems.
Thing is that those systems cost, and you have to make sure that the capacity of that system is the same as the normal park you have. A back-up system normally only have records of the data, it isn't however playable and doesn't have the server load capacity to handle the traffic,It's apparent you're not familiar with how infrastructure virtualization works, I'm not talking about "fail safe" systems wrapped around redundant hardware, I'm talking about infrastructure that can be replicated in software and that is scalable on demand (see VMware, Citrix, OpenStack, etc.., etc,,,). Most business that don't have the heads completely back in the last decade use such technologies not only to optimize their hardware CapEx but also to minimize planned and unplanned downtime on their customer facing systems.it's a pure back-up system, nothing more. As I said, they would have to different server parks with the same capacity on the fail-safe system as on their normal system to not have a server-down situation, and that simply isn't viable in terms of cost.
They have - as all enterprises have - a network setup with back-up systems to take care of the files. This server however is build for that purpose and wouldn't have the capacity to handle the traffic of the players wanting to play.That's not a problem, you patch/fix/restart your primaries while your backups service your clients, once your done with validation you fail back to your primaries (in the case of an MMO you can't do a stateful failback if you're requiring client side patches but it still means minimal downtime to your customers) ...... .Another problem is that some changes need a complete server restart for them to work. This is because the error isn't server based but client based and would require the client to be patched in order for the error to fix itself.Only if your "technicians" don't know what they're doing.....So the best way for the technicians working on the servers are to take down the servers the time it takes to fix it.MMO's only get away with it because the industry has managed to set customer expectations so low, in any other business the amount of downtime these jokers get away isn't tolerated (nor should it be), it isn't necessary with current off the shelf technology.Other MMOs do that all the time, and most - if not all - games with an online feature takes their servers down for maintenance now and then. Most have the schedule to do so with the least interruption, but others might not have that luxury.
mips_winnt wrote: »It's apparent you're not familiar with how infrastructure virtualization works, I'm not talking about "fail safe" systems wrapped around redundant hardware, I'm talking about infrastructure that can be replicated in software and that is scalable on demand (see VMware, Citrix, OpenStack, etc.., etc,,,). Most business that don't have the heads completely back in the last decade use such technologies not only to optimize their hardware CapEx but also to minimize planned and unplanned downtime on their customer facing systems.
That's not a problem, you patch/fix/restart your primaries while your backups service your clients, once your done with validation you fail back to your primaries (in the case of an MMO you can't do a stateful failback if you're requiring client side patches but it still means minimal downtime to your customers) ...... .
BETAOPTICS wrote: »GW 2 truly showed how it should be done though and I miss how good their practices were but I can not ask or expect companies to yet realize that there actually are way better ways to do things.
Posivatedb14_ESO wrote: »I'm just curious on what other peoples opinions are on having a maintenance every week. We pay our monthly fee of 15$ and are not able to play 4-10 (Sometimes more or less) hours a week due to maintenance. I would think in 2014 we would be beyond having to bring a server down for hours at time, and at a time that conflicts with some peoples schedules every single week. I know this topic has come up much in other MMO forums, but I was hoping ESO would be different. I think of this like having a car that you pay for every month that requires you to not use it for a day every week of its existence. What are your thoughts on this? I am also well aware this won't change anything, I'm just interested to see what other's think.
Posivatedb14_ESO wrote: »I'm just curious on what other peoples opinions are on having a maintenance every week. We pay our monthly fee of 15$ and are not able to play 4-10 (Sometimes more or less) hours a week due to maintenance. I would think in 2014 we would be beyond having to bring a server down for hours at time, and at a time that conflicts with some peoples schedules every single week. I know this topic has come up much in other MMO forums, but I was hoping ESO would be different. I think of this like having a car that you pay for every month that requires you to not use it for a day every week of its existence. What are your thoughts on this? I am also well aware this won't change anything, I'm just interested to see what other's think.
Posivatedb14_ESO wrote: »I'm just curious on what other peoples opinions are on having a maintenance every week. We pay our monthly fee of 15$ and are not able to play 4-10 (Sometimes more or less) hours a week due to maintenance. I would think in 2014 we would be beyond having to bring a server down for hours at time, and at a time that conflicts with some peoples schedules every single week. I know this topic has come up much in other MMO forums, but I was hoping ESO would be different. I think of this like having a car that you pay for every month that requires you to not use it for a day every week of its existence. What are your thoughts on this? I am also well aware this won't change anything, I'm just interested to see what other's think.
Wow. This topic actually happened. In 2014. Seriously?