These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

EVE General Discussion

 
  • Topic is locked indefinitely.
Previous page123
 

TiDi in Amarr

First post
Author
Previous Topic Next Topic
Adunh Slavy
#41 - 2013-05-28 12:55:43 UTC
James Amril-Kesh wrote:

That was over four and a half years ago...



Yep, but the basic topology has not changed much as far as I know. I suggested the devblog because a number of people were attempting to stroke their IT epeens.

Necessity is the plea for every infringement of human freedom. It is the argument of tyrants; it is the creed of slaves.  - William Pitt
Delen Ormand
Center for Advanced Studies
Gallente Federation
#42 - 2013-05-28 13:12:32 UTC
Doc Fury wrote:
CCP Veritas wrote:
Doc Fury wrote:
The fact that you guys don't already have the other trade hubs on dedicated hardware is laughable. Dodixie and Amarr especially so.

I submit this in reply:
Doc Fury wrote:
Since you are guessing, please get back to me when you are 100% certain and thus capable to comment.



Then please enlighten us with the specifics of your server pron as relates to this specific issue, and how you have our backs in terms of resource allocation.


I fully agree with Doc here. As a paying customer, I think I have the right to ask some questions when there's the possibility of there being a problem that may probably be allegedly concerning stuff. However, in my case, I would like to know if you use coloured wires on your servers. If that's the case, I would like you to report on what proportion of cabling is colour coded. While you may think this unreasonable, I suspect that any potential time dilation problems are correlated to the ratio of coloured to non-coloured cabling and I would like to be reassured that CCP is going to commit to replacing all standard white or grey cabling in the future in order to sort out the issues.
Ruskarn Andedare
Lion Investments
#43 - 2013-05-28 14:03:07 UTC
Amseln deBrabant wrote:
CCP Veritas wrote:

In any case, Amarr likely should be on its own node, but the loadbalancer isn't doing that right now. We'll likely need to correct that.


just an info, since two weeks or so, there seem to be randomly systems on TIDi. Sometimes with not more than 20 ppl in local and nothing happening. it happend in high sec and low sec as well (Gallente space with FW low sec systems in Black Rise).


That happens due to other systems on the same server being busy.
Feyd Rautha Harkonnen
Doomheim
#44 - 2013-05-28 16:49:55 UTC
CCP Veritas wrote:
We might need to give the load balancer some *subtle hints* if this continues. I'll keep an eye on it.


At which point will techniques similar to 'hourglassing' no longer be used to mask infrastructure performance issues, and the need for a dynamic resource allocating architecture be pursued within CCP?

You can do lipstick-on-a-pig UI stuff like TiDi to hide server performance issues, 24-hour notice to have a re-enforce node set up, or gate-jump animations (which are oh so sexy) to hide the slowness in jumping from system to system as subs increase...but at which point is CCP going to bite the bullet and and fix the core issue of an archicture that isn't truly dynamic resource allocating?

So far I am seeing just UI gimmicks (like the Windows hourglass and Mac spinning rainbow wheel to keep the user entertained visually) rather than fixing root infra scaling issues?

Would you like to know more?
Darth Kilth
KarmaFleet
Goonswarm Federation
#45 - 2013-05-28 17:43:49 UTC
Feyd Rautha Harkonnen wrote:
CCP Veritas wrote:
We might need to give the load balancer some *subtle hints* if this continues. I'll keep an eye on it.


At which point will techniques similar to 'hourglassing' no longer be used to mask infrastructure performance issues, and the need for a dynamic resource allocating architecture be pursued within CCP?

You can do lipstick-on-a-pig UI stuff like TiDi to hide server performance issues, 24-hour notice to have a re-enforce node set up, or gate-jump animations (which are oh so sexy) to hide the slowness in jumping from system to system as subs increase...but at which point is CCP going to bite the bullet and and fix the core issue of an archicture that isn't truly dynamic resource allocating?

So far I am seeing just UI gimmicks (like the Windows hourglass and Mac spinning rainbow wheel to keep the user entertained visually) rather than fixing root infra scaling issues?


I'm pretty sure they got a team dedicated to working on it, the fact you're not seeing immediate results doesn't mean they are just sitting in the park enjoying the weather while drinking some beer doing nothing.

And TiDi is not lipstick-on-a-pig, it masivly decreases the ammount of crashes, black screens and disconnects casued by sudden server load.
Sure it's a bit of drying the floor with the plumbing still broken but it decreases the many issues people had untill they can properly fix and improve the servers and their code.

the 24 notice is because they can only reinforce the nodes by off lining the system, transferring it to the better server and then onlining it again.
Which can only be done during downtime.
Occasionally you see them off lining the surrounding systems and throwing those on separate hardware quickly so the busy system gets the entire server for itself.

And the Gate animations are done by a diffrent team, tough they decrease loading times as a side effect.


They are fixing the code and improving their servers, just because you don't see results now doesn't mean they aren't doing anything at all.
Previous page123