| Author |
Thread Statistics | Show CCP posts - 13 post(s) |
Allen Ramses
Caldari
Typo Corp
   |
Posted - 2009.06.04 20:58:00 -
[1]
 Originally by: Ankhesentapemkah
All nice that you're aware of it and working on it, but we've heard that since October 2008.
Without a single update since. |
Allen Ramses
Caldari
Typo Corp
|
Posted - 2009.06.08 19:41:00 -
[2]
One thing I didn't realize until I saw the HPC/Microsoft presentation was that StacklessIO was a networking layer, and it appeared the proxy blades were significantly reduced. And considering there were endless unordered packet and missing packet errors in logserver when the lag fests occurred, I'd say StacklessIO or the reduction of proxy blades might have something to do with it.
CCP Mindstar, if you're reading, might you be able to rule this out? You are the networking guy after all. |
Allen Ramses
Caldari
Typo Corp
|
Posted - 2009.06.11 20:00:00 -
[3]
Logservers are useless now, the problem has been verified as far as I'm concerned. Every logserver file is flooded with unordered packets and missing packets during combat. It is obviously a networking issue, considering StacklessIO and the reduction of proxy blades were both deployed at around the time we started seeing problems.
So have you been looking at the network saturation or stability? 8 months is long enough to brush off the issue, answer the god damned question, CCP. |
Allen Ramses
Caldari
Typo Corp
|
Posted - 2009.06.11 20:15:00 -
[4]
The problem IS the cause. Laggy FW battles are a by-product of that problem, which is insufficient network allocation. |
Allen Ramses
Caldari
Typo Corp
|
Posted - 2009.06.11 23:06:00 -
[5]
There is nothing to consider. EVE voice runs on a different process, and is not handled by the TQ cluster (AFAIK).
____________________
CCP: Catering to the cowards of a cold, harsh universe since November, 2006. |
Allen Ramses
Caldari
Typo Corp
|
Posted - 2009.06.12 19:50:00 -
[6]
Originally by: CCP Valar
Around 3 weeks ago we replaced the other half of our sol servers. The new servers have 3.3 GHz Wolfdale CPUs and 16 GB of RAM and replace our old 2.8 GHz AMDs that have 4 GB of RAM. With this upgrade we were able to start running all of our cluster on 64-bit processes, but we had to run 32-bit processes on machines that had 4 GB RAM due to the RAM usage overhead of 64-bit processes.
Well, this rules out the geriatric nodes that may have been hosting FW areas. Performance was no different two months ago than it is now. One variable is officially gone.
So, CCP, when are you going to check out the network efficiency? Perhaps removing about half of the proxy blades was not such a good idea? Perhaps your StacklessIO ****ed up your packet scheduling (please tell me you have at least some form of packet scheduling)? If what I know about StacklessIO is true, it was developed in-house, and has had little (if any) testing outside of EVE.
____________________
CCP: Catering to the cowards of a cold, harsh universe since November, 2006. |
Allen Ramses
Caldari
Typo Corp
|
Posted - 2009.06.16 06:10:00 -
[7]
Originally by: CCP Wrangler
We have been testing this on Singularity and we have some leads, currently we're in the process of looking into the data we've gathered and following up on the aforementioned leads.
*Facepalm*
Fine. I'm done trying to get ANY information out of CCP. |
| |
|