oh plz enough already

[ashikenshin]

Just curious. What is as soon as possible about a maintenance scheduled 24-48 hours in advance? That sounds like the opposite of as soon as possible to me, but that's just me.

that's as soon as possible they may know the exploit but they have to program the solution. They also have to test this solution and see if it doesn't break something else. This is not magic you know?

[Rivienne]

If your playtime or timezone puts you far out of these maintenance windows then you really have no weight in the argument. Most of the recent downtown has started between 9-11pm PST, far from a time period where everyone should be sleeping, no?

I am not sure why you felt the need to respond to me at all, but I am fairly sure you completely missed my point since I was not arguing one way or the other on what SE are doing. Merely the nature of the back and forth complaining in this thread being due to the way in which the issues in threads like this are being presented and discussed. Venting will rarely result in constructive dialogue.

Also, not sure why you think my being hit by maintenance (and yes, I have been, more than once) has any bearing on my comment one way or the other. Especially since, again, I was merely commenting on the way in which people were presenting their complaints/requests, not stating one way or another my stance of the maintenance as they are concurring.

None of these numerous bi-daily maintenance's are fixing exploiting, or game and economy breaking issues.. If they were, they would bring the servers down after a 10 minute notice and resolve any major issues of that nature. However, they are scheduling these for 24-48 hours in advance., so it's nothing that severe.

Yes, we all want bugs fixed, yes they have to do maintenance for that to happen. No on is arguing against that.

The issue is why do they need to do that four times a week, instead of one scheduled time slot (like nearly every other MMO on the market)? The same amount of fixes are going to happen when they do one maintenance a week, or seven. In the former case they are just bundling more per maintenance.

Rambling response here. You again seem to have taken my comment out of context: I was not saying SE was doing it right or wrong. I was simply stating that situationally speaking there are valid reasons case by case that could explain what they are doing, and that they might be doing it right, but without real data we can't know.

Basically, you seem to have a slight flaw in your assumption regarding what constitutes an emergency and why. While on the whole you are correct, in that for general maintenance it should be possible to schedule once a week without issue, you seem to have a skewed understanding of how emergency and sometimes non-immediate emergency patches work in the real world.

Sometimes an emergency isn't about immediate problems, but about spiraling problems where after a certain amount of time a small issue becomes a huge issue. You can know you have a certain amount of time to fix the issue before it causes long term problems, and also know if you wait too long it will be harder to control. Sometimes this means truly immediate: everything must come down right now. Sometimes this means you have a cushion of a few hours, or a few days.

For example(this entirely speculative, just to try and provide a tangible examle): if SE found out that there was a currency exploit, they would rarely immediately take down all worlds until they fixed the exploit because of how long it could take.They couldn't simply lock out the economy either like they locked out turn 5, because it is an integral part of the game. So what would they do? They would say nothing. Meanwhile they would monitor the people who made use of the exploit, follow the resulting money at least in scope, and work on a fix. Once they had a fix, they would announce a reasonable amount of time in advance that a maintenance would be coming giving them time to finish QA'ing the fix before rolling it out, and giving players a chance to know they won't be able to get in during that period.

Once they took down the servers they would make the fix. Ban the exploiters, remove the exploited money, and bring everything back up. In some circumstances they might wait until the next scheduled maintenance, but the longer the exploit is in place, the harder it becomes to track the money and remove it without damaging the economy further. So they have to decide: do it now, or do it later. If there was already a delay due to the time needed to implement a fix then depending the severity, they would take it down sooner rather than later. That is how this kind of fix works in the real world. You seem to think "24 hours advance" means it isn't an emergency, but in fact there are varying degrees of an emergency, and scope can be determined by something like critical mass: say, if we leave this exploit in place at the current rate the economy will collapse in 4 days, so we have 3 days max to fix this, but the next scheduled maintenance is in 5 days so we will aim for 2 days.

This also applies to technical flaws, such as index corruption. If you know that a certain behavior is causing index corruption, but you cannot predict when the corruption will occur, you have to monitor the situation and attempt to preemptively do maintenance. Sometimes this could mean a repair on a rolling schedule, such as twice a week, vs once a week. Sometimes it could mean once a week, but not on a fixed day of the week if it is dependent on load etc so you might only have 1 day advance notice. And this will continue until an actual patch that prevents the problem from occurring entirely is released, which might be soon, or might not, depending on the complexity of the issue.

So, yes in fact, sometimes depending on the nature and scope of the issue they can't just "fix it immediately" nor will they wait until the next scheduled day since the longer they wait, the worse the result, maybe even either server crashes, or economic crashes.

Again I am not saying SE is doing it right or wrong, merely that there are many valid reason why they could be doing it right given specific circumstances, just as there are reasons they might in fact be doing it wrong. Without more information we have no certain way of knowing one way or the other. We can only speculate. Therefore without hard data, I am not going to say one way or another that they are or are not handling each individual case correctly. I simply don't know. They could be dealing with issues along the lines of my examples, or they could be dealing with a constant string of bad mistakes and poor planning. Knowing from actual experiences that both can occur, I remain objectively neutral on the question.

[whilke]

that's as soon as possible they may know the exploit but they have to program the solution. They also have to test this solution and see if it doesn't break something else. This is not magic you know?

No software development company is scheduling maintenance for a specific time until the new fix/build/patch is ready to deploy. What if during testing of the fix they found a new use case that doesn't fix the issue? What if when deploying the fix to their staging environment they find out it causes a much bigger issue then when it was just in QA? What if it just took development longer then expected to find and resolve the issue? Often times a bug is going to go back to development to continue fixing until QA is satisfied that the issue is fixed, and no regressions are introduced.

When they are scheduling these emergency maintenance's they have whatever they want to do ready to go (could be OS updates, could be bug fixes, could be physical hardware/network changes/whatever).

And no, it's not magic. It's software engineering at it's basic level.