|
|
| Author |
Thread Statistics | Show CCP posts - 5 post(s) |
Ban Doga
   |
Posted - 2010.02.04 15:06:00 -
[1]
Since now you know (or are getting close to knowing) what caused the nodes to show reduced performance compared to Apocrypha will you be able to prevent this and/or similar things from happening in future releases?
Will this knowledge have any long-term effect in preventing new lag instead of fighting lag after it becomes evident?
|
Ban Doga
|
Posted - 2010.02.08 07:12:00 -
[2]
 Originally by: LaVista Vista
I think it's great to see that it is being worked on. 
If CCP working on it is "great" what would be just "okay" - Not working on it but saying "we know it exists"?
Would not doing or saying anything then be a "bad" thing?
But keep in mind:
The tests to sort this out started only a month ago.
Don't get ahead of yourself and expect them to come up with fix right now.
This is serious business and there is no magic wand or switch they can use to resolve this...
/sarcasm
|
Ban Doga
|
Posted - 2010.02.08 11:19:00 -
[3]
Originally by: CCP Atlas
So, despite 'making people faint' comments this change was done in an extremely controlled manner and was not a case of 'keeping the node alive longer than it should have been'. If the node would have been allowed to die the results would have been extremely bad for everyone...
*confused*
In one sentence you are saying that you did not keep the node alive and then the next sentence starts with "If the node would have been allowed to die"??
So did you prevent the death of the node or not?
Keeping something alive means it would die on it's own.
If it wouldn't die on its own there is no need to keep it alive by doing anything, no?
If this was not a case of "keeping the node alive longer than it should have been" what case of "keeping it alive" was it?
Is this all just a case of poor wording?
What did you really do and what would have happened without it?
|
Ban Doga
|
Posted - 2010.02.08 11:26:00 -
[4]
Originally by: Future Mutant
You ppl whine way too much.
You whine when they dont release details.
You whine when they do release details.
You whine when the node crashes.
You whine when they prevent the node from crashing.
Anyone else noticing a pattern here?
Im pretty much convinced everone living in null is a 4 yr old girl.
Well CCP is not the first company to say one thing but do another.
They are not the first to tell their customers there is no problem although there is one.
They are not the first to start telling self-contradicting stories about what's happening.
Also they wouldn't be the first company to stop doing that.
That's nothing new at all.
But what do you find so admirable about that kind of behavior that you expect everyone to be content about it?
|
Ban Doga
|
Posted - 2010.02.08 23:51:00 -
[5]
Originally by: CCP Atlas
Originally by: BlackHorizon
...from what I can read from CCP Atlas' posts, CCP forced the node to stay online.
That is actually not what happened. A bug which caused nodes to die was fixed. It is as simple as that.
There seems to be some very subtle meaning to this "keeping a node alive".
Apparently it's not even close to what everyone thinks it means...
|
Ban Doga
|
Posted - 2010.02.10 00:29:00 -
[6]
Edited by: Ban Doga on 10/02/2010 00:31:02
Originally by: Mah Kraah
generaly the devs are right when saying that lag and loadissues are random.
in the fight we talk about here, -a- entered system 3+h bevore the fight to avoid that system and grid load problem.
" as if someone told them exactly how to play around the problems(accusing ccp of taking sides)"
well we did 2 weeks of intensive tests in 49 with stressed out servers and several nodecrashes just a few days bevore that d-g fight, thats why we knew what to avoid.
"the attacker know when he attack the defender dont know when to defend so defender is screwed"
BS, the dominion sov system gives the defender 4 timers. he know 4 times in a row, more than a day ahead, exactly at what minute he has to defend and he has the luxory to screw up 3 times and if he does everything right at only one of the timers he has won already.
cva knew 2 days in advance when to be in system and when to force -a-to jump into them.
cva where even in system and had already loaded system and grid but jumped out of system to jump back in again, exactly that was the moment they removed the randomness from the problem: one party loaded one party not. not the code, not the dev in system , cva s decission made the problem a onesided one.
jumping in 150 caps at once and jumping in the in-system support fleet to the same location resuts in many thousend database read and write acessrequests landing in a que on a cpu that is already far beyond 100% usage...... they had to load system and than grid while -a- had to load grid only and than started to shoot and added to the problem by doing so.
If it was truely random nothing they did before the battle should matter. That's the very definition of random: to be unpredictable.
When what you say is true, then the results are somewhat predictable and thus NOT random - it would make that a kind of de facto game mechanics that certain actions cause certain problems in the node and that those problems are likely to strike pilots doing certain actions and likely render them helpless.
Of course we could argue whether this would make it an exploit or not, but I doubt CCP will ever call it that.
But the actual point is that it is possible to create conditions in which players can be killed without them being able to do anything about it. You have to be very naive to think that this can be in the interest of any player.
Because everybody/-thing in EVE tells you
"There is someone who knows more / fights harder / learns faster / has more luck / etc. - If you can win today someone else can win tomorrow"
Oh and a CPU cannot really be far beyond 100% usage - not even a little bit...
|
Ban Doga
|
Posted - 2010.02.13 06:22:00 -
[7]
Originally by: CCP Atlas
Edited by: CCP Atlas on 08/02/2010 12:33:58
Thank you all for your comments. I'll just make a brief comment about the concerns people have about the node being kept alive for the 1600 person fight that is mentioned in the blog.
We are deploying a fix to Tranquility this week (probably tomorrow) which identical to the methods used that fateful night in keeping the node up.
...
We will be deploying fixes this week and the next and are continuing to closely monitor fights as they occur. Please bear with us a little while longer.
Did this really happen?
Since it's saturday I assume there will be no more deployments this week.
This sounds like pure server-side patches so this could have happened silently without anyone noticing.
But wouldn't someone (you?) post that patches are now working and improving the situation for all nodes and players??
|
Ban Doga
|
Posted - 2010.02.13 19:22:00 -
[8]
Originally by: Future Mutant
Originally by: Aralis
Originally by: Future Mutant
Or maybe we can just stop being complete morons about it. Case and point- whose idea was a 1500+ fleet fight again?
CCPs. Pay attention.
I must have missed the ccp post saying "when you attack a system make sure you bring 800+ pilots- because this way you crash our node and can come ***** and moan at us in the forums"
If anyone can find that post feel free to link it.
Maybe you can find the link where it says "no more than 1000 people per battle or our nodes will crash and you will lose ships that we won't reimburse..."
Much obliged
|
Ban Doga
|
Posted - 2010.02.14 14:58:00 -
[9]
Originally by: Future Mutant
Edited by: Future Mutant on 13/02/2010 19:48:46
Originally by: Ban Doga
Maybe you can find the link where it says "no more than 1000 people per battle or our nodes will crash and you will lose ships that we won't reimburse..."
Much obliged
You really need everything spelled out in one place?
Did you read your last posting while writing it?
Just in case you forgot...
Originally by: Future Mutant
I must have missed the ccp post saying "when you attack a system make sure you bring 800+ pilots- because this way you crash our node and can come ***** and moan at us in the forums"
Looks like you need everything spelled out in one place, too.
No idea why you expect people to follow some "guidelines" that aren't written down and not follow some other "guidelines" that aren't written down either.
And when did we start to assume EVE is broken and cannot handle anything unless someone explicitly says "it's okay to do it"?
Originally by: Future Mutant
The cluster cannot gracefully handle a 1600 person fleet fight right now
You realize the linked posting is not even a week old and the problem exists for over 2 months.
Time travel is still not a widely used technology... http://en.wikipedia.org/wiki/Causality
|
Ban Doga
|
Posted - 2010.02.14 21:45:00 -
[10]
Edited by: Ban Doga on 14/02/2010 21:45:51
Originally by: Future Mutant
Under your definition then eve has been and always will be broken. It cant hold everyone in a single system- must be unplayable amirite?
Well wrong- it is playable. There are limits to everything and eve is no exception. It doesnt take a genius to figure out that under the current mechanics the defenders have a rather large advantage. That coupled with the node limitations greatly limits attacking- not defending as its been suggested.
It also doesnt take a rocket scientist to realize that if jita, the most heavily reinforced node, can only handle around 1300 pp then more then that actually fighting is prolly a bad idea.
If ccp somehow managed to get reinforced nodes to support 750 v 750 figts- one of you would find a way to break the node just so you can come back to the forum and complain about how bad things are.
I'm quite amazed about all the things you read into what I wrote.
I never said EVE was broken.
I was asking when we started to asume it is broken,
because your argumentation was "no one said 1500 people in a system will work - so of course it won't work".
Now guess what: no one said 500 people in a system will work, not even 10 vs 10 is mentioned as "will work".
But somehow it is completely obvious - at least to you - that one is okay and the other is not.
Why don't you tell us the magic number where okay becomes not okay - seems like you have a way of determining which number of people in a system is alright and which is not.
|
|
|
Ban Doga
|
Posted - 2010.02.14 21:54:00 -
[11]
Originally by: Future Mutant
The magic number is.... x-1 with x being the amount of ppl in a system that borked the node.
Unless some idiot fired a bomb at the cyno and then its (y/3)* x^3
You're making more and more sense with every post you add to this thread.
Keep going...
|
Ban Doga
|
Posted - 2010.02.18 08:10:00 -
[12]
Originally by: Future Mutant
mentioning qr caused lag and then asking what did they learn? WTF are they supposed to learn? that coding is hard?
[sarcasm]
Of course not. They already know that.
And you are absolutely right. There is nothing to learn from the problems you encountered in the past and the solutions you came up with.
Developing software is basically repeating the magic tantra of "it's difficult - bugs happen - we will fix all bugs that make it into the release" and then throwing in the occasional prayer that everything will work out fine.
There is no such thing as detecting problems you had to deal with in the past and developing strategies to avoid them.
Everybody knows that bugs happen and there's nothing you can do but release the software and see what happens.
Anybody who thinks you can simulate even a small fraction of what's happening on the production system on a test system is obviously not thinking straight.
[/sarcasm]
|
Ban Doga
|
Posted - 2010.02.18 17:19:00 -
[13]
Originally by: Future Mutant
If you want less bugs upon release of a patch- cross your fingers. Bugs happen. You can test for as many things as you can think of but infinity is a big number- good luck expecting perfection.
Again you are explaining there is no way to get less bugs (crossing your fingers is really doing nothing) because there is no way to eliminate all bugs.
There is a difference between those two.
It's the same with "learning from past problems".
It cannot and will not prevent all future problems, but it can prevent having the same problems again and again and make it less likely to have similar problems.
No one demanded perfection.
In fact you are the one using impossible perfection as an argument to justify introducing similar problems again and again.
|
Ban Doga
|
Posted - 2010.06.18 23:53:00 -
[14]
Reading the dev-blog again makes me think you were/are seriously trolling us.
You tell how you gained great understanding and insight and fixes are ready and will be deployed over the next weeks.
Only that this was over 4 months ago and that there are still hilarious stories about whole fleets dieing to failed grids and suddenly re-appearing ships.
The situation is basically unchanged and the only thing you achieved with that blog is stretching your credibility (almost) beyond believe.
|
|
|
| |
|