Pages: 1 2 [3] 4 5 :: one page |
|
Author |
Thread Statistics | Show CCP posts - 5 post(s) |
Hakaru Ishiwara
Minmatar Republic Military School
|
Posted - 2009.12.27 13:54:00 -
[61]
Originally by: Marc Folk
Originally by: Etien Aldragoran blaaa
have you ever scaled large clusters for more than 1000 people? it's hard to decide such a large test... you'll never be able to log a large fleet fight if you ar e not aware of it cause nobody fills the fleet fight form... i'll be their for testing :)
And what if CCP had their logging triggered by something as simple as system population numbers, cyno fields or some other indicator that something big is happening or about to take place? CCP already captures all of this data and publishes it out to the players via the star map.
Of course, the question must be asked: where on earth is CCP QA with their automated testing scripts and resulting logs to ferret the results from changed code with the dominion patch? I can only assume that EVE must be a complex beast to understand and subsequently develop a comprehensive testing regimen, but the concepts of "QA" and "Testing" are not new to the development world even if they are seemingly foreign to the "Agile" methodology.
|
Doc Fury
|
Posted - 2009.12.27 17:38:00 -
[62]
So CCP, how's that Infiniband upgrade working out for you?
The accumulated filth of all their sex and murder will foam up about their waists and all the ho's and politicians will look up and shout 'Save us!' and I'll look down, and whisper 'no.' |
fuxinos
Caldari Guys 0f Sarcasm
|
Posted - 2009.12.27 19:30:00 -
[63]
Originally by: CCP Tanis
Low-sec FacWar tests:
|
Dregek
Pilots Of Honour Aeternus.
|
Posted - 2009.12.27 22:36:00 -
[64]
Edited by: Dregek on 27/12/2009 22:38:05
Originally by: Raider Zero
DON T WASTE UR AND OUR TIME LOOKING FOR SOMETHING THAT NOTHING HAS TO DEAL WITH DOMINION CODE.
I HAD THE SAME LAG AND BUGS ISSUES FROM TPAR ASSAULT (PREDOMINION) THAN GEMINATE ASSAULT
I DID NOT HAVE THE SAME LAG AND BUG ISSUES FROM SEPTEMBER AND OCTOBER GOON AND COMPANY OFFENSIVE IN DETORID.
I M 99% SURE THERE IS SOMETHING WRONG IN THE LAST PREDOMINION PATCHES.
Could you be more ret@rded, firstly writing in capitals doesn't make you stand out it makes you look like a dumb whining 12yr old & they are testing both pre/post dominion code if it shows up in the pre code then they look back further so stfu and be constructive next time
edit: also maybe it was just you! because i didn't have it and people i know didn't have it when we helped IT take pnq
|
Your Host
|
Posted - 2009.12.28 03:22:00 -
[65]
I'm willing to bet isk that the new lag is from the loot logger.
|
Rakshasa Taisab
Caldari Sane Industries Inc. Initiative Mercenaries
|
Posted - 2009.12.28 03:39:00 -
[66]
Probably should also make sure the fleet tests are done with Loot Logging turned both on and off.
|
Doorsdown
Minmatar hirr Morsus Mihi
|
Posted - 2009.12.28 04:43:00 -
[67]
i have noticed huge lag even being generated by shifting wing and squad commanders around. Not to say that the sol cause but the new fleet app seems to be at least one of the problems.
|
Ardetia
The Flying Tigers United Front Alliance
|
Posted - 2009.12.28 11:03:00 -
[68]
All the things introduced in dominion obviously contributed to this problem is, CCP isnt telling us where the fault is, is it CLIENTSIDE or SERVERSIDE ??
CCP coders never had the guts to join 0.0 alliances in battle just to get rid of the problem or maybe theyre not allowed however, considering just how _____ up fleet battles are right now, i dont see why they shouldnt make an exception
i say put them in fleets and let them log in for hours on end to empty grid, and eventually respawn in station
The Flying Tigers are recruiting! |
LUKEC
Destructive Influence IT Alliance
|
Posted - 2009.12.28 11:16:00 -
[69]
Originally by: Ardetia All the things introduced in dominion obviously contributed to this problem is, CCP isnt telling us where the fault is, is it CLIENTSIDE or SERVERSIDE ??
CCP coders never had the guts to join 0.0 alliances in battle just to get rid of the problem or maybe theyre not allowed however, considering just how _____ up fleet battles are right now, i dont see why they shouldnt make an exception
i say put them in fleets and let them log in for hours on end to empty grid, and eventually respawn in station
That's not nearly horrid enough. Let them mine their ships first, then go fight fleet battles (though recently, it's been okish to me)..
Why don't you make new gallente titan event or something like that ?
|
Urraka
|
Posted - 2009.12.28 11:58:00 -
[70]
Originally by: CCP Tanis Howdy folks, Test schedule:
Tuesday [January 5] @ 17:00 GMT (0.0 Fleet/Sov tests) Thursday [January 7] @ 17:00 GMT (Low-sec FacWar tests) _________________________________
Considering a good majority of EVE players actually have a semi sort of life and do go to work. Wouldn't a slighty later hour be appropriate? 17:00 GMT that's the time most people get off work in EU and lunchtime in Major parts of the Mid and East coast US.
|
|
Seishi Maru
The Black Dawn Gang
|
Posted - 2009.12.28 14:05:00 -
[71]
I bet 3 isk it is related to fleet finder :)
|
Aenachreon
|
Posted - 2009.12.28 19:47:00 -
[72]
Originally by: Seishi Maru I bet 3 isk it is related to fleet finder :)
The very first fleet engagement I was in after Dominion, I joined a large fleet while docked and was promptly lagged out with the fleet window open. Once I closed that, everything was fine.
|
Sawyer LaFleur
|
Posted - 2009.12.29 04:43:00 -
[73]
Its got to be a top priority for you guys to be able to get some sort of client server conversation going on to keep a ship from uncloaking or appearing on grid after a jump, warp, station exit or other state change or to at least allow you to postmortem the situation adeuately.
I think we're also owed a reporting of numbers of ships which were non responding in a sytem at given times and have some broad info shared if the outages had any geographical or operating system, or game settings corelations.
Nothing exact, but share what you do know and what you're working on to bread confidence.
|
SeismicForce
Terra Incognita Systematic-Chaos
|
Posted - 2009.12.29 10:18:00 -
[74]
CCP should completly remove the fleet function (if possible, or just the bonus) for the tests. I suspect a lot of lag comes from the calculations required when jumping into a system, as every bonus each pilot gets, has to be calculated per squad, then per wing, then per fleet. If a bonus-giving member crashes then the whole lot must be recalculated. Just my 2 isk.
Originally by: Allisie In a recent interview, a dev mentioned that ships and skills cause lag and will be removed in EVE 2.
|
CyberGh0st
Minmatar Ara Veritas
|
Posted - 2009.12.29 13:39:00 -
[75]
Originally by: Lijhal Edited by: Lijhal on 25/12/2009 15:55:24 you want to fight the lag ? here are 2 (maybe 3) solutions!
1) stop pumping server resources into jita for all those 4-4 huggers who are playing 0.01 isk wars and scam the daylong jita local ... just pointless
That remark was the dumbest one I've read today :p
And because you probably won't be able to figure it out on your own :
1. It is pretty logical to reinforce Jita, because the load is permanent. 2. It is only one system, so the actual hardware that goes in there is peanuts in comparison with the complete cluster. 3. Reinforcing and improving Jita gives CCP valuable live information and experience to improve other systems.
For the trolls : I haven't been in Jita for months.
http://www.mmodata.net Favorite MMO's : DAoC SI-Era / SWG Pre-CU-NGE |
Hakaru Ishiwara
Minmatar Republic Military School
|
Posted - 2009.12.29 14:12:00 -
[76]
Originally by: SeismicForce CCP should completly remove the fleet function (if possible, or just the bonus) for the tests. I suspect a lot of lag comes from the calculations required when jumping into a system, as every bonus each pilot gets, has to be calculated per squad, then per wing, then per fleet. If a bonus-giving member crashes then the whole lot must be recalculated. Just my 2 isk.
Oh, but wouldn't it be in CCP's best interest if that infernal fleet window is open? Damn thing causes a 75 - 90% "FPS" decrease in my client every time there is an update to the fleet personnel. Let's not make this test easy for our CCP QA friends.
For testing purposes, the window is open and displaying the hierarchical fleet view. Even with all wings minimized, this local client performance hit takes place. If the "history" (aka broadcast) tab is selected, then this performance hit disappears.
|
Zeerover
DeadSpace Exploration and Investigations
|
Posted - 2009.12.29 22:04:00 -
[77]
Originally by: thisisnotmikaldrey
Originally by: Ban Doga Your standing can and will become 10.0 (even without skill bonuses).
And there are more than just 5 variants. E.g. anchoring a POS in highsec needs 10x the faction standing of the system you want to anchor in. So your -10 -5 0 +5 +10 would require this to be changed.
let me clarify . .
you dont actually get +10 it will show +10 but its 9.9999999999 and the +10 is actually rounding as the standings only actually show 10.00 (2x decimal places.)
Dev Link or stop attempting to clarify something you don't know about:)
When you reach multiple 10.0 standings you'll notice that they're not ordered in numerical order, but rather alphabetical order, since the game considers all the standings equal, and this alphabetical order includes everyone hard coded 10.0 starter agent.
|
Praetor Novak
Macabre Votum Morsus Mihi
|
Posted - 2009.12.30 00:57:00 -
[78]
In addition to the usual lag reducing steps I've noticed that keeping a streamlined overview, closing as many chat windows as possible and closing the fleet window helps this new Dominion lag monster.
|
Hakaru Ishiwara
Minmatar Republic Military School
|
Posted - 2009.12.30 18:02:00 -
[79]
Originally by: Praetor Novak In addition to the usual lag reducing steps I've noticed that keeping a streamlined overview, closing as many chat windows as possible and closing the fleet window helps this new Dominion lag monster.
Good points.
The irony here is that in order for the game to be playable in larger-scale fights, the game must be pared down to its basic elements with the removal of most graphical effects, useful overview columns, chat windows that may provide additional intel beyond what is stated on voice communications and a fleet window that also now houses fleet member broadcasts.
And with Apocrypha's removal of the ability to right click in space and align / warp-to a celestial or bookmark from the system map (without actually clicking the celestial in the map), we can not even feasibly fight using the map view as a last-resort lag reduction method.
What a pickle.
|
SFX Bladerunner
Minmatar Black Serpent Technologies R.A.G.E
|
Posted - 2010.01.01 17:15:00 -
[80]
I will try to be there with 2 accounts. my plane lands around 6:30am (GMT+1) on tuesday so should be enough time for me to get back home and get set up.
What I would like to know though is:
Will CCP assign some sort of FC's to the (I assume) 2 fleets fighting eachother or will we be able to 'select' our own FCs?
I think fighting without FCs would be silly because everyone would just shoot whoever they deem a good target (pick a hostile ship thats nicely in your optimal and can be taken down by yourself). If you want to recreate lag issues properly I think proper target calling is in order (and thus making everyone in fleet A shoot a few primaries and vice versa.
Might seem a bit over-the-top but who knows. It's more realistical and it might just help CCP find+solve the problem __________________________________________________
History is much like an endless waltz, the three beats of war, peace and revolution continue on forever.. |
|
Kaurapa
|
Posted - 2010.01.03 03:41:00 -
[81]
So after making us pay for a terrible system that they lauded as being a break through - CCP now want us to help fix it for them.
Give me 1 plex for each character I bring and I'll be there.
|
zythyl
Gallente ANZAC ALLIANCE IT Alliance
|
Posted - 2010.01.03 05:27:00 -
[82]
Edited by: zythyl on 03/01/2010 05:29:15
Originally by: Kaurapa So after making us pay for a terrible system that they lauded as being a break through - CCP now want us to help fix it for them.
Give me 1 plex for each character I bring and I'll be there.
How about stop turding around and help make the game better?
By your logic, patches are a way of CCP saying: "Hey look at our revolutionary system. Now wait while we fix it, mkays? Might take a while... I dunno... maybe five freaking years?"
EVE is always being worked on, always upgrading, improving over time. The devs said there will only ever be one EVE. And they intend to make it last "TILL THE END OF TIME". I for one will be sending all my accounts. The lag is, I admit pretty bad. But that's just because *some people* decide it's a good idea to warp into a 600popped node without informing CCP.
|
Kaurapa
|
Posted - 2010.01.03 05:43:00 -
[83]
Originally by: zythyl
EVE is always being worked on, always upgrading, improving over time. The devs said there will only ever be one EVE. And they intend to make it last "TILL THE END OF TIME". I for one will be sending all my accounts. The lag is, I admit pretty bad. But that's just because *some people* decide it's a good idea to warp into a 600popped node without informing CCP.
Great advertising by a ccp employee?
Re your "some people" comment: CCP were informed. Cant reinforce nodes until DT under present systems. So the timers created by CCP for use on TCUs mean that to stop a corp/alliance from gaining sov you have to fight with out reinforcing the node.
The mistake that "some people" made was being too stupid to realise what they were attempting to do had been done by others less than 2 weeks ago. No surprise the results are the same.
|
Obsidian Hawk
Free Galactic Enterprises FREGE
|
Posted - 2010.01.03 07:59:00 -
[84]
I still hope CCP tanis takes my thoughts into consideration
The fleet battle tests need to take place at a gate and a pos tower. NOT a random beacon
We need to have hot drops.
We need ccp in titans shooting other capitals.
We need to have a bubble party too. Everything and everything that goes into a fleet battle needs to happen.
|
ceaon
Gallente
|
Posted - 2010.01.03 17:32:00 -
[85]
do we get any free subscription days for helping CCP whit this thing ? helping for the sake of common good is not very profitable |
Ol' Delsai
Caldari Kernel of War Tau Ceti Federation
|
Posted - 2010.01.03 18:04:00 -
[86]
Edited by: Ol'' Delsai on 03/01/2010 18:07:44 Edited by: Ol'' Delsai on 03/01/2010 18:06:10 Since it was lost in an ocean of trolls , here's some facts/solutions to the current lag monster issue that happens wether or not the node is reinforced ... It seems to be a grid issue, not a node issue
This should be reproduced during the test on sisi ...
Thanks to Jarnis McPieksu who originally posted this on CAOD
Quote: Dear EVE players who haven't figured this out yet, here is a...
Guide to Dominion Grid Load Lag
Post-Dominion, we got a new bug. NC experienced it first-hand a couple of times and lost a pile of ships to it. Through repeated attempts at bashing a skull to a wall, a solution was eventually found. Many fleet battleships died bringing you this information.
If you try to enter system (not just warp to a grid, enter a new system) on a grid that has too many other ships on it, you will end up in a bugged state - grid will not load for you but your ship appears there and will most likely get shot - "too many" is a number that is somewhere between 200-300 ships. You also will be unable to log back in, so CTRL+Q is not really helpful.
Key points;
- It does not matter how many people are in system. It does not matter if the system is reinforced or not. The one and only thing that matters is how many people are on the grid you are using to enter the system. - The number matters at the moment you try to load the new system and grid and even blues count.
Hence, if you try to jump through a gate, if the other side has a large fleet, you will most likely get completely hosed. However, a mass jump can also cause everything to break for a part of your fleet simply because they enter the system only after the gate grid has already "too many" friendly ships. Cue "first 100-200 people got thru fine, rest are lagged out on load, and look, reds are now arriving on grid and... *welp*".
If you cyno in, same is true for the grid where the cyno is.
Hence, for the love of all that is holy etc., do not light multiple cynos on a single grid and do not try to squeeze more than 200 ships through at once (number of ships jumping in + number of ships already on grid).
The way to combat the problem is to stagger the entry, use multiple grids, clear out ships from the grid as soon as they land to reduce problems for anyone still jumping in and preferably use cynos that are so far away from any hostiles that they cannot land on the cyno grid in time with a blob and break your entry before you are through and have warped off the cyno grid.
After you have properly entered the system, then everything works just like pre-dominion. 700-800 ships, while laggy, is playable. 1000+ could work on a reinforced node. Entering grids full of ships in that same system, while they may take a while - 10-20 seconds is not unheard of - will work. Cue fun mega bash without one side winning with a lame turkey shoot.
Until CCP can fix their code, hope this helps.
|
Tzujeih
GoonFleet GoonSwarm
|
Posted - 2010.01.03 19:25:00 -
[87]
Your inability to adequately simulate player connections is not encouraging.
|
Your Host
|
Posted - 2010.01.03 23:26:00 -
[88]
plz sticky this thread in all the forums, not just general! and give it that pretty pink outline so people will notice it. and maybe put it in the client at our login screens. maybe even put it in the news. SPAM SPAM SPAM!
I'm tired of this 'Spatial Distortion'
|
Zenst
Aliastra
|
Posted - 2010.01.04 09:18:00 -
[89]
100 or even 500+ is not fleet battle testing now is it :(.
Seriously - why havnt CCP written a test harness to test this stuff out, do it once, record peoples action were you can then reply them to the server - thats how other people test server/client systems for robustness. WHen testing a http server do they get several hundred people to conenct to that web server, dont think so. How about you drag yoruselfs into the 90's on that at least.
Also until you start sorting out the whole log area you can expect little empathy for such testing. How hard would it be to have a client version of logs thats encrypted with a public key that only CCP can decrypt with a private key, that would be easy fix to the whole we cant except client log area and avoid the bull of outr log show nothing, yet at same time they dont show a valid action that leads to the end result you reaising the ruddy petition for.
So in summry:
Sort out a test harness that can test this or you will get what we pay for kind of results given the huge efforts needed in setting up test clients that then get updated the next day forcing you to do it all again. Whilst at it perhaps drag your bug system into reality as currently its about as useful as a shared notepad file that nobody else can read even when they have to write to it (sorry hope i didn;t devuldge inner working there, just observant).
Sort out client logs you can use - disraputed computing is not just processing offloading you know, think about it. Not exactly effort to do this in a simple/secure way.
If you did that instead of breaking things every ruddy time and then crying help then you only indicate that the live server is unable to log things at a level you can fix things with.
Also explain what the issue is - is it one big process loop thats spawns threads with a half-life that means most timeout before actioning once a certain level of users get in a system. Its a resource issue and were is that bottleneck - I/O thread starvation, beyond sloppy code were non the wiser.
Help us to help you and then when we do help, sort these very very simple things out I have outlined: test harness(record/replay clients) and trusted client logs, realy isn't major issue and I'd say a good programmer could knock up these in a very very short time, honestly.
|
Sargine West
|
Posted - 2010.01.04 10:36:00 -
[90]
I just tried to finish 2 clients for tomorrows testing but run into a few issues.
Any new info on the Multiplicity patch for tomorrows testing, i'm afraid there is currently none, so i cant patch. Also SISI is running a different version than the TQ build, is this going change?
|
|
|
|
|
Pages: 1 2 [3] 4 5 :: one page |
First page | Previous page | Next page | Last page |