Pages: [1] 2 :: one page |
Author |
Thread Statistics | Show CCP posts - 0 post(s) |

Rohann
|
Posted - 2006.03.04 00:16:00 -
[1]
20k plus players online at peak times, one server and lots of problems. And this posted on the patch notes page does not give me confidence with CCP's dream of a single shard anymore.
Quote: A number of fixes and code improvements have been made to reduce/eliminate stuck issues. These include: Checking whether a system is stuck/loaded prior to your being allowed to undock or jump into it (if it is, then your autopilot will deactivate and leave you at the gate). Checking to see if the system is loaded when trying to log in just after a reboot (if you logged in space). The map has a "travel advisories" option that will display systems that are known to be currently stuck. In all cases you will receive an explanatory error message.
To be able to see if systems are stuck by looking at the map only confirms what I have been saying all along. This server cannot handle it. What do you think?
|

Golerre Evraun
|
Posted - 2006.03.04 00:19:00 -
[2]
um... bad piece of code... serverside memory leak... node memory fills....node lags...gate closes. ctd. ---------------------------------------- If you ain't dyin, You Ain't tryin. |

Antoinette Civari
|
Posted - 2006.03.04 00:20:00 -
[3]
Originally by: Rohann To be able to see if systems are stuck by looking at the map only confirms what I have been saying all along. This server cannot handle it. What do you think?
The server can handle the amount of players and systems, but not the memory leak which seems to exist.
|

Nafri
|
Posted - 2006.03.04 00:22:00 -
[4]
Originally by: Antoinette Civari
Originally by: Rohann To be able to see if systems are stuck by looking at the map only confirms what I have been saying all along. This server cannot handle it. What do you think?
The server can handle the amount of players and systems, but not the memory leak which seems to exist.
wonder if its the same, the client has to suffer
|

Teblin
|
Posted - 2006.03.04 00:22:00 -
[5]
The server can handle the load it is currently recieving, however there is an unrelated memory leak issue which is causing the lag we are experiencing. Please see kieron's sticky for information about patch and server problems.
|

MysticNZ
|
Posted - 2006.03.04 00:27:00 -
[6]
I love people like you. -
                        You got pwnd by us too :P - Wrangler lol - Imaran |

KirikaPhoenix
|
Posted - 2006.03.04 00:29:00 -
[7]
As has been said, its fairly simple :P
Server leaks memory Node strain increases Node implements Traffic Control to try and reduce strain, keep itself alive Systems shut down Repeat until - A) Server collapses B) Dev comes along, shouts 'Bad Server!', whacks it a couple of times with a *insert heavy blunt instrument of choice* and manually resets congestion system, clears memory leak.
Give them some time to fix the memory leak, they couldn't have spotted it during testing since it only seems to occur with 10k + players present. Trying to debug the server whilst its falling over isn't terribly easy 
|

Jinx Barker
|
Posted - 2006.03.04 00:31:00 -
[8]
Can someone explain the "Memory Leak" to me? Is that something to do with RAM? And how it fills Up fast or something? Sorry, I am trying to be technical here... 
|

KirikaPhoenix
|
Posted - 2006.03.04 00:32:00 -
[9]
Memory leak is where theres a bug in the program that causes it to ask for more memory for whatever reason. Usually its some form of loop that shouldn't happen e.g. it doesn't have an exit clause.
Program continues to ask for more and more and more memory even though it doesn't need it. It never stops until its exhausts all the avaliable memory in the system and causes it to crash out.
|

Tauruz
|
Posted - 2006.03.04 00:33:00 -
[10]
Edited by: Tauruz on 04/03/2006 00:34:39 I remember when EVE broke 7000 simultaneous players; people said the server wouldn't be able to handle any more.
I think 'the server' can handle it, I just think CCP are trying too hard to prevent people whining on the forums when they get stuck on heavy load evenings (eg Sundays). Thats why we now have 'travel advisory' code!
|

ArcticFox
|
Posted - 2006.03.04 00:33:00 -
[11]
Everyone else who's posting is correct, this is not an issue of the server being unable to handle the number of players.
I'd also like to point out that sharding the game would indeed solve CCPs load problem: a huge chunk of the Eve community would simply quit. ---------------------------------- "There's no +6 Sword of WTFPWN in Eve." - Er... Some person on the forum... |

Mind
|
Posted - 2006.03.04 00:34:00 -
[12]
Originally by: Jinx Barker Can someone explain the "Memory Leak" to me? Is that something to do with RAM? And how it fills Up fast or something? Sorry, I am trying to be technical here... 
Memory leak is when a application (ie, EVE server software) allocates memory but never free it up again, and so the program slowly fills the memory.
this can be casued by a realy tiny issue, i bet the code has ALOT of places where it alocates memory for diffrent things (keeping track of people, transferes, killed npc's etc) and one tiny one is all that needs to fill the RAM.
|

Rohann
|
Posted - 2006.03.04 00:45:00 -
[13]
Edited by: Rohann on 04/03/2006 00:47:30
Originally by: Teblin The server can handle the load it is currently recieving, however there is an unrelated memory leak issue which is causing the lag we are experiencing. Please see kieron's sticky for information about patch and server problems.
Yeah I have read that however the problems with the server have been gradually increasing since Exodus, as with the player count. The best patch CCP ever came out with was CASTOR. Hell im still waiting for another "Castor" patch.
I do not have much knowledge in server structure or programming or even networking. With that said all of my opinions about the server come from gaming experience. I have done a test with my bookmarks to see if I could reduce client lag. What I did was I took all of my bookmarks out of ppl and places and put them in secure cans so that I could retrieve them later. So now I am running a client with NO BOOKMARKS whatsoever. The client runs ace. Hardly any combat lag when figthing in 0.0.
My proposistion:
Code it so that when you warp from gate to gate you actually land on top of the gate. This would force players to delete their bookmarks which in turn would increase gaming performance in combat. It would also speed up traveling time for those that do not have bookmarks currently. Its been a highly debatable subject with pros and cons. However I think we have to think of paying customers gaming experience in this case rather than that of game mechanics. I feel this is a very good idea for the game. Now I know this wont make the server run better but it will help the game a bit.
Now if you will excuse me I have to put all my bookmarks back in my ppl and places so I dont get pwnd this weekend in my battleship.
|

Capsicum
|
Posted - 2006.03.04 00:47:00 -
[14]
Wikipedia has a good article on memory leaks too.
forum rules | [email protected] | Our new Website! |

Wild Rho
|
Posted - 2006.03.04 03:23:00 -
[15]
Originally by: KirikaPhoenix As has been said, its fairly simple :P
Server leaks memory Node strain increases Node implements Traffic Control to try and reduce strain, keep itself alive Systems shut down Repeat until - A) Server collapses B) Dev comes along, shouts 'Bad Server!', whacks it a couple of times with a *insert heavy blunt instrument of choice* and manually resets congestion system, clears memory leak.
Give them some time to fix the memory leak, they couldn't have spotted it during testing since it only seems to occur with 10k + players present. Trying to debug the server whilst its falling over isn't terribly easy 
That's about the best explanation I've ever seen 
|

xeom
|
Posted - 2006.03.04 03:34:00 -
[16]
I think I see a post with this exact same name every patch.
It is a pipe dream....coming true everyday. ---------------------- O Snapz! |

Gen Maton
|
Posted - 2006.03.04 04:22:00 -
[17]
Originally by: KirikaPhoenix As has been said, its fairly simple :P
Server leaks memory Node strain increases Node implements Traffic Control to try and reduce strain, keep itself alive Systems shut down Repeat until - A) Server collapses B) Dev comes along, shouts 'Bad Server!', whacks it a couple of times with a *insert heavy blunt instrument of choice* and manually resets congestion system, clears memory leak.
Give them some time to fix the memory leak, they couldn't have spotted it during testing since it only seems to occur with 10k + players present. Trying to debug the server whilst its falling over isn't terribly easy 
B) is better known as percussive maintenance or impact therapy.
|

JamesTalon
|
Posted - 2006.03.04 04:39:00 -
[18]
You could call it shock therapy, but only since the server would be shocked as hell when that solid object came slamming into it  
Caldari Navy Surplus
Eris was here ~Eris |

Danton Marcellus
|
Posted - 2006.03.04 05:17:00 -
[19]
I just wanted to be on a whine thread.
Remind me about The Maze.
I'm Danton Marcellus and I approve of this message. |

Livia Tarquina
|
Posted - 2006.03.04 05:19:00 -
[20]
Who hasn't put new hardware and software into their computer and had problems. Even experienced people run into problems building their own computer. This is what happened here so I give them a break.
"Big guns and heavy armor what else is there?"
--Amarrian Admiral before entering battle against Jove Navy |

Budz Fergie
|
Posted - 2006.03.04 08:24:00 -
[21]
There is a memory leak client side too. After playing for a couple of hours I notice that both Eve accounts are using 400-500megs or ram each and this started right after the RMR patch. A reboot clears it up, but, it keeps repeating.
|

Corp Scammer
|
Posted - 2006.03.04 08:36:00 -
[22]
something as complex as eve is bound to have some bugs the devs have some powerful database tools it all works out CCP requested feature please - quite simple really.
For CEOs days since last login details on all members in their Corp. this will help manage inactives and those that dont log in can be removed - |

Jon Baptist
|
Posted - 2006.03.04 08:38:00 -
[23]
Just out of interst, how much physical RAM do you have? And are you sure it's eve that's not freeing the memory up? I'm running my client here for long stretches of time (I have no life) and I haven't seen any degredation in client speed that I would associate with a Memory leak.
Try leaving your task manager on without eve running, but running any other applications (TS or IRC or Whatever) for a couple of hours and see if this duplicates the problem.
|

F'nog
|
Posted - 2006.03.04 08:53:00 -
[24]
Originally by: Gen Maton
Originally by: KirikaPhoenix As has been said, its fairly simple :P
Server leaks memory Node strain increases Node implements Traffic Control to try and reduce strain, keep itself alive Systems shut down Repeat until - A) Server collapses B) Dev comes along, shouts 'Bad Server!', whacks it a couple of times with a *insert heavy blunt instrument of choice* and manually resets congestion system, clears memory leak.
Give them some time to fix the memory leak, they couldn't have spotted it during testing since it only seems to occur with 10k + players present. Trying to debug the server whilst its falling over isn't terribly easy 
B) is better known as percussive maintenance or impact therapy.
It's Impact Maintenance. But I kinda like Percussive Therapy better. I think I'll use that from now on.
Now where's my mallet...?
Originally by: Bl4zer But, cmon, this is the Eve forums, we don't let facts get in the way of pointless speculation.
|

Rafein
|
Posted - 2006.03.04 08:56:00 -
[25]
TBH. the server was working pretty good after the hardware upgrade, until this patch. But it seems a memory leak slipped through in the Blood patch
|

Jenny Spitfire
|
Posted - 2006.03.04 08:57:00 -
[26]
Originally by: Rafein TBH. the server was working pretty good after the hardware upgrade, until this patch. But it seems a memory leak slipped through in the Blood patch
A little drops of water makes the ocean?  ----------------
RecruitMe@NOINT! |

Tremain
|
Posted - 2006.03.04 09:38:00 -
[27]
Sharding wouldn't help - some parts of Eve are universal - your character, the map, the market is regional.
Everything else is per system, so in effect every system is a shard and your character is transferred between the shards when you jump.
This is more complicated that the usual MMO idea of having a little server with a couple of hundred people all handled on one CPU but I certainly wouldn't say it is a pipe dream - infact I'd say it was a pipe dream to think sharding would help in any way.
|

Vladimir Ilych
|
Posted - 2006.03.04 09:47:00 -
[28]
Have faith in CCP. They will fix the issue and a multiple shard is bad. m'kay?
|

Dors Venabily
|
Posted - 2006.03.04 10:25:00 -
[29]
I have full confiddence they will
|

Jim McGregor
|
Posted - 2006.03.04 11:24:00 -
[30]
I love the fact that ccp is at least telling us what the problem is. In many other mmo's, customers wouldnt have been told the technical reason behind the problem since i guess they think its bad publicity or underestimate the customers ability to understand.
--- "Automatic override. Manual control overridden by autopilot. Please wait for operation to complete. You can override the automatic autopilot override in 28 seconds. Then you can make it wait" |
|
|
Pages: [1] 2 :: one page |
First page | Previous page | Next page | Last page |