The point of a test is to test.
There are several existing issues that CCP is looking to resolve and this test will help narrow down solutions.
Yes, the respawn of many systems (explo, asteroids, older missions, etc.) in the game is critical, and one of the specific issues that this test will help better understand. There is a lot of code running EVE and some of it has limited documentation.
This test is not about getting rid of downtime.
Pretty sure they were shooting straight this time.
It’s an experiment to see what needs to be looked at later.
—Gadget likes a bit more rock smiting time, but worries about a lack of rocks…
Here’s the thing. They are broken up onto many regional servers. So they can do it at locally quiet times.
Anyhoo: Why is it being done on live?
Because a test on a system which has 20+ K users connected at a time, is quite different from one which gets a couple of hundred. Things build up differently.
You know CCP have been working to reduce downtime for years? It used to be around an hour. It’s now less than 5 minutes (when they’re not doing deployments)
It’s a sucky user experience to be kicked off, even for only 5 minutes.
It’s even suckier to lose a kill because of it.
The amount of salt about this is baffling to me.
They are doing a test. Even if they mirror the entire Tranquility cluster to Singularity, this does not give them the load that the production system experiences.
I don’t think they have ever had more than 1000 - 2000 people on any “Mass Test” on Singularity so that is not a practical argument.
They have made it very clear this is a ONE DAY experiment to show them where they need to focus their efforts.
The long term outcome of a test like this is the potential for us to have less downtime over time. This is a good thing.
Not every post from CCP requires a “world is ending” flame post barrage… only most of them LOLOL 
well there is fundamental question…
what will be done with asteroid belts? lets say we get in future one downtime per week,so week of no ore in asteroid belts? or they gonna make belts autorefresh?
With no downtime to kick all the afk players off I wonder if this will lead to a proper AFK timeout?
wow… the doomsayers are out in full force today. How could anyone see this as a bad thing? Sure, by thursday at 10 utc things might be a bit laggyer than normal… but eve has never been exactly lag free. even on the best of days it takes a full second between pressing F1 and my guns shooting.
This is an experiment. IF things go poorly maybe we can all get some free plex out of it. If things go well maybe we will have fewer and fewer downtimes in the future.
Source for this claim?
Where did CCP say it’ll happen “randomly” ?
That can be done by putting in a mechanism to respawn asteroid belts or the like. It’s like with World of Warcraft where instances (i.e. ateroid belts in EVE) can be respawned without taking the whole lot down. Databases don’t have to go offline (depends on what they use as DB) and refreshing cache can be done with a cronjob.
I do like this. Thumbs up 
. Hope it works
It’s easy to guess the reason for this experiment. The Koreans have started to play EVE Online and the current downtime slot is right at the time the Koreans start logging in for their evening.
I don’t see the asteroid belts becoming the big problem. Of course most belts will be empty Wednesday evening because they haven’t been refreshed, but that can be solved in other ways with simple respawn mechanics.
It’s the mentioned memory leak and caching issues that I think are the real problem. I suspect the server processes taking more and more memory the longer they are running and only a reboot of those processes gives the server memory breathing room again. And because there has always been a daily reboot, fixing those memory leaks were not that high prio. But now, when they would actually be forced to fix them, in 20 year old code, that can be a huge PITA.
We want to gauge the impact to inform our priorities. What should we fix to try and get to a place where we can normally run Tranquility for, say, 48 hours at a time and perhaps have four downtimes each week. And then continue to increase the run-time and reduce the downtimes by focusing on the pain points as they surface.
What I’d like to know is why is it so important to remove the daily 5 minute DT?
I haven’t seen any complaints about the regular DT nor have I seen any major outcry for it’s removal.
I’m sure there’s a lot of other more important pressing issues in-game that should be addressed and fixed instead of working towards removing DT.
My preference, tbh, is to kill off asteroid belts entirely, and require a degree of scanning to find anything decent.
But that’s a major mechanics update, so… 
The cheap and nasty solution is: Lots more anomalies. With some regular ore anomalies added in.
You don’t know many Australians, do you?
Where can I find out what sort of server park they use and OS?
This. 1000x this.
Make exploration a real job / opportunity in EVE.
1 massive asteroid belt per system (like a real asteroid belt. MASSIVE). 98% junk. 2% Asteroid Ore, Moon Goo, PI, Etc.
You have to scan the belt to find pockets of ore. These are not default instant warpable.
People who take the time to scan them down can sell the locations.
Conversely the miners will be safer because they will no longer be at a click and warp location. Hostiles will have to scan them down.
SOOO many opportunities for a better eve in this one idea alone. I have been advocating for this for years.
This would be ideal. More ore anomalies and zero belts. Same with ice.
A jet can disappears in a set amount of time. Other resources can too.
I pleased with this change. Downtime has become part of the meta, and it shouldn’t be.
In addition to no downtime, randomizing the downtime within a one or two hour window so that it can’t be relied upon as a tool for structure timing would also go a long way towards dealing with this issue while still allowing the benefit of downtime in terms of fixing memory leaks and all the back end stuff.
Probably last major update put into public. But data is probable bit dated.