Daily server downtime and rollbacks

[The Stormbringer]

Is the game fixed yet? Should I wait a few months before buying it?

[Vespian]

Is the game fixed yet? Should I wait a few months before buying it?

The game is great, but there are peak times, depending on your geographical location, where the queues can hurt. It will be fixed, they've released a statement recently, that they are starting a PTR for the coming changes.

Crashes still occur, but very minimally (yes, you can die during such a crash). Apart from that, I have no issues with the game.

[The Stormbringer]

Cool, thanks! Maybe I'll pick it up near Christmas.

[Zeta333]

When a small indie company releases a brand new game like this, you can expect a few years of bugs.

The ironic part is the ONLY fucking part of the game blizzard had to manage was the servers, the ONLY part. and look what fucking happened.

- - - Updated - - -

The game is great, but there are peak times, depending on your geographical location, where the queues can hurt. It will be fixed, they've released a statement recently, that they are starting a PTR for the coming changes.

Crashes still occur, but very minimally (yes, you can die during such a crash). Apart from that, I have no issues with the game.

Its not peak depending on where you are. Its when its asias primetime there is a que everywhere. And on the weekends usually a que from small to big all weekend. Its cus asia flooders thier realm then floods europe and americas.

I ran ubers from keys to torch and crashed 5 times to desktop.

[Vespian]

The ironic part is the ONLY fucking part of the game blizzard had to manage was the servers, the ONLY part. and look what fucking happened.

- - - Updated - - -



Its not peak depending on where you are. Its when its asias primetime there is a que everywhere. And on the weekends usually a que from small to big all weekend. Its cus asia flooders thier realm then floods europe and americas.

I ran ubers from keys to torch and crashed 5 times to desktop.

I think the ubers are currently still a known crash area, but throughout the game, you're pretty safe to run 8 hours straight, pretty much.

[Zeta333]

I think the ubers are currently still a known crash area, but throughout the game, you're pretty safe to run 8 hours straight, pretty much.

I didnt crash in uber trist. I crashed in all other areas. taking waypoints, just getting an army for my summons ect

[Jotaux]

When I do ubers I have someone else in the game as a safety net.

[EyelessCrow]

The game is great, but there are peak times, depending on your geographical location, where the queues can hurt. It will be fixed, they've released a statement recently, that they are starting a PTR for the coming changes.

Crashes still occur, but very minimally (yes, you can die during such a crash). Apart from that, I have no issues with the game.

Honestly whenever I've tried, regardless of time of day there has been a long queue. Also crashes are NOT minimal, at least not for me. Every 3rd or 4th game is a crash. If I walk away for an hour I come back to a crash log screen. This seems to be the common experience if you look anywhere instead of covering your eyes and ears going "The game is great"

[Vespian]

Well, I can't help you guys, I only get crashes once every so many hours, on average. I've done about 300 Shenk & Pindle runs the past few hours, only 2 crashes. No queues.

- - - Updated - - -

The ironic part is the ONLY fucking part of the game blizzard had to manage was the servers, the ONLY part. and look what fucking happened.

Also, I missed this one, but why are you back to blaming the servers? They have nothing to do with this and we've already proved it in this very thread, to you.

[Zeta333]

Well, I can't help you guys, I only get crashes once every so many hours, on average. I've done about 300 Shenk & Pindle runs the past few hours, only 2 crashes. No queues.

- - - Updated - - -



Also, I missed this one, but why are you back to blaming the servers? They have nothing to do with this and we've already proved it in this very thread, to you.

I mean first we already know what level player you are but the fact your only running shenk and pindle. the servers are entirely the problem. Legacy code is legacy code but that is solvable by expanding server capacity and throughput. Lets not even fucking pretend blizzard doesnt have extra hardware or the ability to add more to solve this issue. But hey we are coming up on big game releases soon and they are about to lose alot of players if they dont pull thier heads out of thier asses.

[kranur]

I mean first we already know what level player you are but the fact your only running shenk and pindle. the servers are entirely the problem. Legacy code is legacy code but that is solvable by expanding server capacity and throughput. Lets not even fucking pretend blizzard doesnt have extra hardware or the ability to add more to solve this issue. But hey we are coming up on big game releases soon and they are about to lose alot of players if they dont pull thier heads out of thier asses.

Well people already paid for D2R so blizz loses nothing ... If they stop playing it's even better for them as that will remove server stress and not force them to upgrade.

[Vespian]

I mean first we already know what level player you are but the fact your only running shenk and pindle. the servers are entirely the problem. Legacy code is legacy code but that is solvable by expanding server capacity and throughput. Lets not even fucking pretend blizzard doesnt have extra hardware or the ability to add more to solve this issue. But hey we are coming up on big game releases soon and they are about to lose alot of players if they dont pull thier heads out of thier asses.

So much to unpack there.

I mean first we already know what level player you are

I have no idea what this is supposed to say, why it would be relevant to the subject of "server capacity". Please elaborate as to what level player you think I am.

but the fact your only running shenk and pindle.

That evening I was only running Shenk and Pindle, because after my sorceress got RIP'd, I didn't feel like rushing another one, while I lack the +skills for one. So I'm basically farming with various characters, slowly moving my other characters through the game in preparation to farm more. The RIP on my sorceress came effectively too soon, which resulted in a major setback on my general progress. Normally I would be farming Baal Hell solo at this time, but, I really don't gaf about your assumptions based on what I decide to run that evening. Also, there is no relevance to the issue. For a Shenk & Pindle run, I transition 3 times through portal/wp's within 1 minute. Your argument was, that doing lots of different things, going from area to area, resulted in crashing. I don't share that experience, I have been running multiple characters through hell, solo, and no crashes for hours. (On PC). Besides, your ass runs SC, so stfu

the servers are entirely the problem. Legacy code is legacy code but that is solvable by expanding server capacity and throughput.

No, it is not solvable by adding hardware. This is clearly your severe lack of understanding of the situation and the underlying technology, I can't even begin to imagine what type of bitchy manager you are IRL, at least one that knows better at all times. The Service is a single point of contact (failure, in this case) for all interactions, globally. So there is basically a process, that process runs as a service, that service is basically a script with functions and that single script is supposed to be able to cope with all interactions of the entire world, processing forth and back all the changes to the region servers. The script happens to be legacy code, unoptimized, probably with no or very limited multithreading and needs to be rewritten. It doesn't matter how many servers they put in that pool, the script remains the single point of contact. The bottleneck, if you will. It will not matter.

And additionally, to add, They seem to have fixed the game creation/join issue, at the same time of rolling out the PTR for future fixes, so even less complaints.

[Zeta333]

So much to unpack there.

I have no idea what this is supposed to say, why it would be relevant to the subject of "server capacity". Please elaborate as to what level player you think I am.


That evening I was only running Shenk and Pindle, because after my sorceress got RIP'd, I didn't feel like rushing another one, while I lack the +skills for one. So I'm basically farming with various characters, slowly moving my other characters through the game in preparation to farm more. The RIP on my sorceress came effectively too soon, which resulted in a major setback on my general progress. Normally I would be farming Baal Hell solo at this time, but, I really don't gaf about your assumptions based on what I decide to run that evening. Also, there is no relevance to the issue. For a Shenk & Pindle run, I transition 3 times through portal/wp's within 1 minute. Your argument was, that doing lots of different things, going from area to area, resulted in crashing. I don't share that experience, I have been running multiple characters through hell, solo, and no crashes for hours. (On PC). Besides, your ass runs SC, so stfu


No, it is not solvable by adding hardware. This is clearly your severe lack of understanding of the situation and the underlying technology, I can't even begin to imagine what type of bitchy manager you are IRL, at least one that knows better at all times. The Service is a single point of contact (failure, in this case) for all interactions, globally. So there is basically a process, that process runs as a service, that service is basically a script with functions and that single script is supposed to be able to cope with all interactions of the entire world, processing forth and back all the changes to the region servers. The script happens to be legacy code, unoptimized, probably with no or very limited multithreading and needs to be rewritten. It doesn't matter how many servers they put in that pool, the script remains the single point of contact. The bottleneck, if you will. It will not matter.

And additionally, to add, They seem to have fixed the game creation/join issue, at the same time of rolling out the PTR for future fixes, so even less complaints.

So your telling me a company that has run a 25million player count at one time MMO and one of the longest lasting with the largest population mmos since 2004 cannot figure out how to fucking increase the server capacity for a game made in 1999 in a month let alone a week. stop making excuses for them. They have the experience, ill bet money they have unused hardware that over the course of a few weeks they could temp retool to handle the fucking load and then repurpose for somthing else when the load dies down. If your excuses is the servers for diablo 2 are so fucking complex that blizzard entertainment with the backing of activision cannot fucking fathom how to increase server cap then i have a bridge to sell you.

They have been fumbling and fucking up literally everything they have touched for the past several years. I dont trust them with diablo 4. I dont trust them with wows future, fuck they couldnt even handle a reskin of warcraft 3 ffs. Thier lack of communications and no one post about the issues every 1-2 weeks shouldnt be acceptable is atrocious.

[Vespian]

So your telling me a company that has run a 25million player count at one time MMO and one of the longest lasting with the largest population mmos since 2004 cannot figure out how to fucking increase the server capacity for a game made in 1999 in a month let alone a week. stop making excuses for them. They have the experience, ill bet money they have unused hardware that over the course of a few weeks they could temp retool to handle the fucking load and then repurpose for somthing else when the load dies down. If your excuses is the servers for diablo 2 are so fucking complex that blizzard entertainment with the backing of activision cannot fucking fathom how to increase server cap then i have a bridge to sell you.

They have been fumbling and fucking up literally everything they have touched for the past several years. I dont trust them with diablo 4. I dont trust them with wows future, fuck they couldnt even handle a reskin of warcraft 3 ffs. Thier lack of communications and no one post about the issues every 1-2 weeks shouldnt be acceptable is atrocious.

No, I'm not telling you anything about server capacity, because server capacity isn't the issue. How you keep coming back to that retarded conclusion is beyond me.

[Zeta333]

No, I'm not telling you anything about server capacity, because server capacity isn't the issue. How you keep coming back to that retarded conclusion is beyond me.

when to many people play at the same time a que is enabled to handle increased traffic. Thats a result of not enough capacity of some sort, wether bandwidth or just plain capacity. Fucking regardless of the issuem ques durring peak asia time for all realms is fucking stupid and easily solveable, so its a que token. Wow has it. So many qol improvements that would enable people to skip ques yet are not there.

[Vespian]

when to many people play at the same time a que is enabled to handle increased traffic. Thats a result of not enough capacity of some sort, wether bandwidth or just plain capacity. Fucking regardless of the issuem ques durring peak asia time for all realms is fucking stupid and easily solveable, so its a que token. Wow has it. So many qol improvements that would enable people to skip ques yet are not there.

They explained it in their statement
We explained their statement to you, in this thread
I just explained their statement to you, again.

I'm sorry, but you'll have to concede that this field of expertise, is not yours.

The queues and the game creation\join delays were raised to reduce the stress on the service, not the servers. I have exhausted my ability to explain to you in layman's terms why it's not primarily a server capacity issue, but you keep going back to that, because you don't understand IT, probably in general. It doesn't make you an idiot to admit that this is simply not a field you have knowledge of, it is however extremely stupid to keep hammering an issue that is not there, when you clearly do not understand the issue at all.

I'll do another attempt.

Have you ever heard of the analogy "9 mothers, 1 baby in 1 month"? It's something we use a lot to explain why adding more, isn't always effective. The idea is that people always immediately go for the obvious solution, which is that if 1 server isn't enough, surely, adding more will reduce any issues that are taking place on that one server. We then refer to the fact that you can't reduce the pregnancy's period by adding more women.

This is such an issue, due to reasons that are far too technical for you, in this particular case, adding more servers isn't a solution.

[EyelessCrow]

when to many people play at the same time a que is enabled to handle increased traffic. Thats a result of not enough capacity of some sort, wether bandwidth or just plain capacity. Fucking regardless of the issuem ques durring peak asia time for all realms is fucking stupid and easily solveable, so its a que token. Wow has it. So many qol improvements that would enable people to skip ques yet are not there.

What Vespian is trying to say is (I think)

The root issues are with database servers writing too frequently, which they seemed to have fixed by stopping regional databases writing to the global database. Adding new database servers wouldn't have fixed that issue. As well as some kind of legacy service repurposed for modern times. The legacy service cannot keep up with the demand they expected to be for Diablo 2:R so they are having to rework the underlying systems because that particular legacy service can only be ran by itself, unlike modern systems that have multiple instances of things like that. So the comparison to more modern titles doesn't really fit.

Networking is my field of expertise not DBA or programming but that's what I could gather from Blizzard's post.

[Biomega]

The legacy service cannot keep up with the demand they expected to be for Diablo 2:R so they are having to rework the underlying systems because that particular legacy service can only be ran by itself, unlike modern systems that have multiple instances of things like that. So the comparison to more modern titles doesn't really fit.

To be fair, part of it is code but part of it is behavior as well. I think what they primarily underestimated might have been THE WAY people play, not how many people would play. The fact that many popular ways of farming involve rapid creation of new games really sets this apart from most games - Pindle runs can easily clock in under 1 minute each, and Mephisto runs can easily be done in a minute or two. That's not the kind of game length you'd see in most games, and it doesn't take many people to do this all day every day to put considerable strain on the database.

[Vespian]

To be fair, part of it is code but part of it is behavior as well. I think what they primarily underestimated might have been THE WAY people play, not how many people would play. The fact that many popular ways of farming involve rapid creation of new games really sets this apart from most games - Pindle runs can easily clock in under 1 minute each, and Mephisto runs can easily be done in a minute or two. That's not the kind of game length you'd see in most games, and it doesn't take many people to do this all day every day to put considerable strain on the database.

But this has been the norm for 2 decades.

So, to be fair, it is a form of capacity, it's just not server capacity. They miscalculated or mistheorized that this single service could process ten, maybe a hundred times the original load and clearly they missed the mark. By their description, the script was originally unoptimized, the database synchronization illogical, and nothing was done to modernize it, which I assume, can only come from a decision made based on an estimation that went wrong.

[EyelessCrow]

To be fair, part of it is code but part of it is behavior as well. I think what they primarily underestimated might have been THE WAY people play, not how many people would play. The fact that many popular ways of farming involve rapid creation of new games really sets this apart from most games - Pindle runs can easily clock in under 1 minute each, and Mephisto runs can easily be done in a minute or two. That's not the kind of game length you'd see in most games, and it doesn't take many people to do this all day every day to put considerable strain on the database.

I'd argue it's entirely the system. Player behavior wouldn't be an issue if they didn't build it off the same system they used back then utilizing a single service.

I don't understand the underlying decision to "preserve legacy systems" that players don't knowingly interact with, the UI is already different. Why was that the line they didn't want to cross? If they hadn't gone that route I feel like they could have accomplished the same thing by building it over a new system, which sounds like they have to do anyway. Again I'm not a programmer so maybe I'm talking out of my ass.