Server Hosting
 

Zodiac and Era need to be hosted on better servers. For the most popular servers on Graal, they sure do lag a lot.

Or they just delete useless Scripts from Levels. :o

I would love to hear from Stefan or Unixmad (or anyone else) on what kind of hardware these servers are running on. It sounds like they have MySQL running on some sort of Linux box, but how powerful are we talking?

Furthermore, do you think if they leveraged more CPU or RAM to these specific servers that the Graal code would even take advantage of it?

I've noticed Zodiac lagging pretty bad lately. I apparently have around 200ms (which is normal for a Euro player) but my spells sometimes have up to 2 seconds of delay before casting.

Phantos <3.

I checked 211ms when I logged on.

Another PW may be indirectly causing the lag on Zodiac.

Talked to Stefan and he mentioned that he would be restarting the servers sometime this week which may help alleviate the lag issues.

Yeah, something tells me that other servers such as UN which have a plethora of unused and messy scripts are inadvertently causing lag on Zodiac... I think it's safe to bet that Zodiac has some of the, if not the very, best scripts currently in use on Graal...

It's not going to be an issue on the classic list more likely would be on an UC server.

2. 0.189685424 % KevinDB (in level onlinestartlocal.nw at pos (30.5, 30))
3. 0.155373811 % *Forums
4. 0.102733771 % *Achievements

these 3 things are making fairly regular SQL queries and never manage to jump up past .2% or so, probably just as likely that someone is doing something taxing in short bursts like /scriptscan all *

If there is such a thing as one server lagging another server due to server A's bad scripts, well, I feel sorry for anyone hosted on the same server as Valikorlia. We have a ton of old systems cobbled together doing who knows what.

Aha! It's Val that is causing Zodiac lag. Delete val!
jkjk

There's obviously such a thing since they're running on the same server and no virtualization seems to be running.

I won't pretend I've ever seen Val's scripts but it seems like a vast majority would be running on the clientside, just from logging on and walking around a bit.

I took a look through NC the other day and we just have dozens of old systems wrapped into one another, nothing really being used heavily, but I feel like some clean up may be in order.

Also when I ran Gball for 3 days straight, I can't imagine that not causing some slow down, flood limit or no.

One of these days I'll get a clearer picture of where and how Graal is hosted and then either be amazed at how little it requires to go, or just how much I've underestimated it all this time.

GBall is only using memory when running, and occasionally a bit of disk IO. You can use memstats (or /stats) to see how much memory it's using.

Interesting stuff in there, but nothing persistently draining, sort of like your previous conclusion.

I know that playerworlds run on Amazon Cloud server hosting, or what ever it's called. Stefan said something like that back when G3D had a team 'working' on it.

They don't yet. Stefan was at one point looking into using Amazon's platform, though. I don't know what came of it.

EC2 wasn't even out of beta until 2008 so I kind of doubt he was talking about it when working on G3D, unless G3D has been worked on more recently than I remember.

EC2?

Amazon Cloud works fairly well for a lot of services. Would be interesting to see them transition over to that. I assume they rent rack space in some network warehouse in Europe, currently.

2009 there was a team working on it. Me and Zero and some others (Vinka, actually)

And if there's a server space somewhere, it should be in the US, seeing as Americans get the better pings.
Btw Phantos since when do you post on here?

A very bad SQL index was using quite a bit of CPU, it's gone now. The query that was causing a problem is now 1000x faster.

Yes we have also restarted all servers to clear up some memory, and will fix the problem with Kingdoms Debug soon.