Ups and Downs - The Server Thread

[lazerdix]

it should be noted that a 1.8ghz server with a gig of ram is plenty hardy enough to support the current use base. the problem is more likely related to a slow internet connection and poorly optimized networking code.

[theTrav]

it should be noted that a 1.8ghz server with a gig of ram is plenty hardy enough to support the current use base. the problem is more likely related to a slow internet connection and poorly optimized networking code.

LOL

1 - People are complaining about crashes. They have nothing to do with speed at all.

2 - YOU are complaining about lag, but you should read Loftars other posts on the matter.

From memory he's got the web server with the BB software and an authentication server all running on the same machine, he also has a big fat chron job that runs at about 7:30pm GMT+10 time every day and produces massive spikes that last for up to a minute.

From memory he has a great fat internet pipe that makes me jealous, and from looking at his network code it's not particularly greedy at all.

[Ruggan]

SOME MIGHT SAY THE SERVER GRINDED TO A HALT

[Harleyvitality]

Others might say it GROUND to a halt.

[Chakravanti]

it should be noted that a 1.8ghz server with a gig of ram is plenty hardy enough to support the current use base. the problem is more likely related to a slow internet connection and poorly optimized networking code.

This. I get the feeling people who are talkign shit about the specs know very little about the difference between PC specs and Server Machine specs. I doubt that without a google they have any clue what the difference between buffered and unbuffered RAM is. User interface on a PC and Remote interface on a game connection are hardly equivilent by a long shot. We'de need an extra hundred ACTIVE players (That's 3x current peak averages) before you start to notice performance lag due to hardware. Having a cutting edge i7 (Or Intels new OctoCore with 32mb caches *Drool*) with 32GB of buffered ram would not solve much lag except maybe those 2:30am EST maintenance spikes. SHit like that can crunch hundereds of thousands of connections per second given the bandwidth. WOuldn't take very many of those to juggle sites liek google or yahoo. Expanding code errors could just as easily cripple, lag, and crash THAT system as the current one.

Besides, Getting Uber PC hardware to compensate for dev stage coding performance issue is the wrong approach anyway. It's like the difference between driving a luxury vehicle and a muscle car. The two are incomparable. One is for old foagies, rich fucks and drug dealers. The other is for a die hard mechanic with a dream (or a pirate). You can't FEEL a luxury car. If this metaphor is lost on you then you really shouldn't participate in this kind of thread anyway.

Also, grammer Nazi's on an international forum FTL.

[Ruggan]

Others might say it GROUND to a halt.

I PREFER GRINDED LIKE LOFTAR DID TO ME IN A CLUB

[Golbez]

Necroposting in order to say: DOWN!

[Lightning2]

The server is up

[GapingAngus]

The server is down

[NaoWhut]

Stilll..... man i g-g-g-gotta get some
of that H&H man! you know anyone
who be willin to hook me up?