open All Channels
seplocked EVE Information Portal
blankseplocked New Dev Blog: TQ Level Up
 
This thread is older than 90 days and has been locked due to inactivity.


 
Pages: first : previous : ... 2 3 4 5 [6] 7 8 9 10 ... : last (11)

Author Topic

CeneUJiti
Posted - 2010.06.17 09:10:00 - [151]
 

Will this even reduce 00 lag and crashes to pre-dominion levels? And has our beloved CCP ever admitted that Dominion made things worse?

Gallion
Amarr
Posted - 2010.06.17 09:21:00 - [152]
 

I Don't Understand all this Jiberwookie IT Talk but it sounds so epic.

I got thinking what if you need to Expand that room for more systems? By that I mean HDD and Stuffs in those Gallente Cabinets. I know its kinda Early for such a Question.

Also I will bow and kiss the feet of all you @ CCP When I see you @ the next Festival.
>hopefully I make it next time<

So the room has Cold Air chilling into the cabinets but why not do something similar on the back area of there as well? (I like the Idea of a Freezer Chilled Computer room).

RedClaws
Amarr
Macabre Votum
Morsus Mihi
Posted - 2010.06.17 09:35:00 - [153]
 

Originally by: Gallion

So the room has Cold Air chilling into the cabinets but why not do something similar on the back area of there as well? (I like the Idea of a Freezer Chilled Computer room).



Walk-in Fridge
Walk-in Fridge 2
Do it yourself


Inquisitor Serinity
Posted - 2010.06.17 10:03:00 - [154]
 

I'm working in the IT-Business and i must say...

I WANT TO TOUCH IT :D

Clansworth
Good Rock Materials
Posted - 2010.06.17 10:22:00 - [155]
 

Originally by: CeneUJiti
Will this even reduce 00 lag and crashes to pre-dominion levels?
Aside from your blatant trolling intent here, let me point out that this is mostly a physical move.. aside from some better networking hardware, there is no real 'upgrade' here. This move on the 23rd is all about preparing for future improvements. The current data center was at a physical limit, electrically, thermally, and physically. the new one appears to have a lot more room for future improvements. I'm guessing the DUST514 servers might start off sharing the same locale (more for support purposes than actual performance concerns), even further demanding the extra space.

As for lag concerns, they have been saying for MANY years that the problem is computational speed, and their single core SOL architecture. Until they determine a safe method to break SOL processes up to multiple cores, LAG will still be proportional to system load. The next improvement appears to be some form of auto node remapping, allowing them to more effectively shift SOL's to different machines to 'reinforce' nodes on the fly. It will be a huge improvement, but not the end-all solution to the lag-beast.

Abelerr
Posted - 2010.06.17 10:36:00 - [156]
 

How far things have come =)

reported by CCP Hellmar | 2003.12.12 00:26:41
"We just completed the quite successful 2 GB memory increase for our SQL Cluster. We see a dramatic change in SQL Cluster performance from our end; I hope you are enjoying the reduced lag."

Liandra Xi
Amarr
The New Era
C0NVICTED
Posted - 2010.06.17 11:15:00 - [157]
 

Quote:
Step three: Pics or it didn't happen


Ooooh shiny! I find myself afflicted by this strange condition called SOS, otherwise known as Shiny Object Syndrome.... I demand moar pictures!@!

Super Whopper
I can Has Cheeseburger
Posted - 2010.06.17 12:33:00 - [158]
 

Edited by: Super Whopper on 17/06/2010 12:35:24
Originally by: CCP Fallout
On Wednesday, June 23, 2010 between 0900 and 1500 UTC, Tranquility will be down while we move the cluster. CCP Yokai's new dev blog gives all the details on the move and improvements to the cluster.

By the way, this is Yokai's first dev blog. Some of you may remember him from Fanfest 2008, when he and a pick up team (Bubblegum!) made it all the way to second place in the mining PvP tournament :D

Update: The blog has been updated with the downtime schedule, which is as follows:

0900: All EVE Services go offline. (Web, Forums, Test Servers, EVE Gate, TQ, basically everything hosted in London)
1200: EVE Online web, secure and Test Servers come back online. (all network services reestablished in London. Only TQ should still be down at this time)
1500: TQ back online



What does EVE have to do with The University of Tennessee at Chattanooga?

Requiem Jofama
NibbleTek
Pandemic Legion
Posted - 2010.06.17 12:36:00 - [159]
 

Edited by: Requiem Jofama on 17/06/2010 12:35:56
Just realized TQ cluster is pretty small compared to high performance computers out there for instance one we use here:
http://www.pdc.kth.se/resources/computers/ekman/ekman-hardware

but at the same time latency is much more important on TQ and hopefully the new server-park will help the hamsters alot.

Bloom Flamestar
Posted - 2010.06.17 12:37:00 - [160]
 

Originally by: CCP Fallout
On Wednesday, June 23, 2010 between 0900 and 1500 UTC, Tranquility will be down while we move the cluster. CCP Yokai's new dev blog gives all the details on the move and improvements to the cluster.

By the way, this is Yokai's first dev blog. Some of you may remember him from Fanfest 2008, when he and a pick up team (Bubblegum!) made it all the way to second place in the mining PvP tournament :D

Update: The blog has been updated with the downtime schedule, which is as follows:


0900: All EVE Services go offline. (Web, Forums, Test Servers, EVE Gate, TQ, basically everything hosted in London)
1200: EVE Online web, secure and Test Servers come back online. (all network services reestablished in London. Only TQ should still be down at this time)
1500: TQ back online




don't recall past times work on the present
1. stable Live Server
2. daily downtime issue should be enough
3. Updates promises should be fulfilled not been deleted
4. new improvements to ships should finally be realized
5. Honesty should be a key Issue in CCP post and not throwing sand into the Eyes of the Community and covering up the mistakes of the Staff

Solomon XI
Hidden Souls
Posted - 2010.06.17 12:41:00 - [161]
 

Clearly we should expect a 24-36 hour period without being able to play EveO. I forsee massive stability issues.

G'Kar5
Gallente
Intaki Research and Manufacturing
Distorted Percepts
Posted - 2010.06.17 13:39:00 - [162]
 

Originally by: Lord XSiV


Well I guess you are validating what I stated to a certain extent but are still way off base if you think Cisco produces good product.


Yes, Cisco does produce a good product. Is it expensive and overpriced, YES. But blaming TQ lag on CCP choosing a Cisco box over Foundary box is just silly.

Originally by: Lord XSiV


Then again you probably aren't or have been, nor will be for matter, in a position to architect a solution of critical significance. This is obvious due to your linkage to lightreading, the basic 'tabloid' equivalent in the tele/datacom space....



Yes, I have helped to build several SP networks over the last 20 years. Please stop attacking me personally.

Originally by: Lord XSiV



Furthermore, Cisco doesn't build their gear, its all outsourced including assembly. They don't even design their ASICs either, that is outsourced as well. As I said earlier they are a software company and that's it. You can drink their 'kool aid' all you want it still doesn't make a difference. Just like their definition of a 'A' class IP address is wrong, but don't tell a CCIE that or he might cry.... So in reality, strike Cisco off your list of vendors knowing 'IP'



Hmmm, not sure what to make of this. Yes, Cisco is mostly a software company, but the router OS is one of the most important things out there. For any given hardware platform, you can make it look great or blow it up; you just have to know its strengths and weaknesses.

Some of their ASICs are designed in house, others are not.

What is a "Class A"? We live in the world of CIDR now.

Originally by: Lord XSiV


I would ramble on more but it would probably be pressing the limits of some NDA points but regardless, it is still considered junk by top tier professionals. The simple and steady rule is you go big iron when important stuff is on the line. These days with the pricing being so low it might be advantageous to go with it anyhow once you factor in TCO. Cisco generally has the highest TCO because you are required to have an expensive support contract in comparison to a less costly support arrangement with Juniper. Plus you have the added knowledge knowing you don't have to worry about hardware failures due to quality issues Cisco is notorious for.



Calling Cisco pricey and calling Cisco junk are too different things entirely. Junk, no; high priced, yes. Granted, Foundary has a very nice product but its Ethernet only. We are still a long way off from getting rid of TDM as some places still only have TDM access.

I'm done now. Have a nice day.

Darek Castigatus
Immortalis Inc.
Shadow Cartel
Posted - 2010.06.17 14:20:00 - [163]
 

so nerdgasms, tech geekery and some mild trolling.

Just a regular dev blog in eve online Wink

Fearless M0F0
Incursion PWNAGE Asc
Posted - 2010.06.17 14:22:00 - [164]
 

Originally by: CCP Fallout

0900: All EVE Services go offline. (Web, Forums, Test Servers, EVE Gate, TQ, basically everything hosted in London)
1200: EVE Online web, secure and Test Servers come back online. (all network services reestablished in London. Only TQ should still be down at this time)
1500: TQ back online



What about THE DEVICE? Shocked

It's important... it does stuff

Lord XSiV
Amarr
Kodar Innovations
Posted - 2010.06.17 14:41:00 - [165]
 

Originally by: G'Kar5


I'm done now. Have a nice day.


Oh so you quit. I win yet again - but not like there was any question as to the outcome in the first place.

Your starring e-peen accomplishment was 'I have helped to build several SP networks over the last 20 years....' in response to my stating you have done nothing of significance. It is a good thing you confirmed it. Low end to mid tier providers are typically ignorant using inferior technology due to the fact that they hire unqualified resources such as yourself.

Now if you had stated something like "I lit up the first oc-768 for commercial use" or was the architect for one of the world's largest financial transaction networks then you might have some creds. But until you do, go crawl back under your rock and keep providing sub par advise to whoever employs you. Kids these days, no respect whatsoever.....

Oh and let me re-iterate, Cisco is junk. Always has been and always will be. It is ingrained into the company culture just as Apple prefers to steal technology rather than develop it on their own. As for your 'router OS' statement, my response it is 'man routed' - it is, has and always will be all you need.

And lastly, lulz at the monkey claiming 20 years 'experience' who doesn't know what a class A IP address is. Do you have your CCIE as well? Did you cry?






Dacil Arandur
Habitual Euthanasia
Pandemic Legion
Posted - 2010.06.17 15:02:00 - [166]
 

Originally by: Lord XSiV
And lastly, lulz at the monkey claiming 20 years 'experience' who doesn't know what a class A IP address is. Do you have your CCIE as well? Did you cry?


I'm pretty sure that part was a joke, since he even specifically mentioned classless routing in it.

Thunder1971
Caldari
NailorTech Industries
RAZOR Alliance
Posted - 2010.06.17 15:23:00 - [167]
 

Well good luck CCP with the whole operation.

Hope it goes smoother then our fleetfightsVery Happy


T

Emrys Ap'Morgravaine
Caldari
Minmatar Ship Construction Services
Ushra'Khan
Posted - 2010.06.17 16:49:00 - [168]
 

You know, as super awesome as all this is, it suddenly occurred to me, that somewhere there's someone (or a group of someone's) who work for IBM or Cisco who are going to be eve players thinking "what awesome bit of kit can I design that'll give me a better experience"

I envy them.

Em.

Commander Azrael
Red Federation
Posted - 2010.06.17 17:06:00 - [169]
 

Originally by: Lord XSiV
Originally by: G'Kar5


I'm done now. Have a nice day.


Oh so you quit. I win yet again - but not like there was any question as to the outcome in the first place.

Your starring e-peen accomplishment was 'I have helped to build several SP networks over the last 20 years....' in response to my stating you have done nothing of significance. It is a good thing you confirmed it. Low end to mid tier providers are typically ignorant using inferior technology due to the fact that they hire unqualified resources such as yourself.

Now if you had stated something like "I lit up the first oc-768 for commercial use" or was the architect for one of the world's largest financial transaction networks then you might have some creds. But until you do, go crawl back under your rock and keep providing sub par advise to whoever employs you. Kids these days, no respect whatsoever.....

Oh and let me re-iterate, Cisco is junk. Always has been and always will be. It is ingrained into the company culture just as Apple prefers to steal technology rather than develop it on their own. As for your 'router OS' statement, my response it is 'man routed' - it is, has and always will be all you need.

And lastly, lulz at the monkey claiming 20 years 'experience' who doesn't know what a class A IP address is. Do you have your CCIE as well? Did you cry?



I bet your an asbolute joy to work with.....

Cisco is NOT junk, expensive yes, but junk no. Yes Brocade, Juniper and alcatel make some fantastic products but using the top end netiron for a LAN is insane. How you can presume to spec an environment when you haven't even seen it? It's also worth pointing out that partnership factors into these kind of architectural decisions, we as a company get huge (40%+) discounts on cisco kit which makes the majority cheaper than the brocade and juniper alternatives and when our customer is paying the bills, it's up to them to sign it off.

We architect, develop, build and support the largest Video on demand platforms in the world from AT&T to disney, C4 to the seesaw platforms and we use cisco almost entirely for these environments because they're the most effective given the specification.

Yeah we have Juniper and Brocade devices on our edge and are migrating to netiron for our edge BGP routing but saying the biggest network equipment vendor is 'junk' is stupid, mis-informed and in all likelihood trolling.

Do you work for brocade or something? Or are you just an advocate of the high port density their devices provide? Cisco suits TQ fine and the minute they need 2 billion packets a second I'm sure they'll look to brocade.

You sound like you just need to smoke a fat one and chill out. Cool

Flame away, as I'm sure your ego wont let these comments go unanswered....

Othran
Brutor Tribe
Posted - 2010.06.17 18:24:00 - [170]
 

Edited by: Othran on 17/06/2010 18:26:02
"While the routers we use are quite powerful (Cisco 7600's with the RSP720 route processors)"

Mmmm you sure you don't work for Cisco? Razz The last time I heard the term "powerful" and 7600 in the same sentence was a few years ago. Been a few years since I was in Triangle Park with the "folks".

Best of luck - there's always something physical that doesn't happen, hope yours is trivial.

Edit - as a sometime contractor for Cisco I can confirm that there is only one "Evil Empire" and its not Intel, MS or (recently) Apple Twisted Evil

CCP Yokai

Posted - 2010.06.17 18:58:00 - [171]
 

CrazzyElk - Is there an approximate guesstimate on the ETA of the next blog.

Answer - Nothing exact, but I am going to go out and say weeks, not months. Just have some things to get done first so we can talk about what we are doing, not what we might do.


Hawk TT - Correct me if I am wrong, but SOL nodes should benefit from Intel QPI, larger & faster caches, more memory, more memory channels & bandwidth, Intel Turbo-Boost feature for single-threaded apps?

Answer - More than likely. Again, something we are in active conversations about with both IBM and Intel.


Hawk TT - Why not "Boot from SAN"?

Answer - Right now the ease of management is not a big issue. It's a fairly small number of servers, the data on the local disk is nothing important, and honestly we are using the high end (lots of MTBF)Hard drives so, failures are few and far between. Not that it isn't easier/better booting from HBA/LUN0 but just quite a few items down the “to do” list.


Vahz Rex - Yokai, you briefly mention management tools in the dev blog, any chance you will cover this more in future blogs?

Answer - Yes, Remapping Eve is really about the tools we have to control what solar systems go where and how to balance. I'll talk to the software guys to see just how deep we can go down that rabbit hole.


Hustomte - Is it possible to get this very important announcement added to the Eve-Gate calendar?

Answer - Great idea... Sent a note off to community just now!


Dacil Arandur - What are the chances something like that could happen for this big move? (webcam for TQ move)

Answer - Sounds like fun, but have you seen us? Probably not the most exciting thing on cam and I couldn't handle the rejection if no one watched.


Lee Dalton - What are the RAM SAN and SSD SAN dedicated to?

Answer - Both are dedicated to the DB


joe hamil - ty for the insight into your end CCP, i will keep on turning up to mass testing as often as i can

Answer - Thank you! I know we always need more people on test servers... and it really does help.


Anikadir - haven't seen a covered cool aisle like that before. Do they make much difference?

Answer - Yes, it does. When you have unmanaged airflow... you really have a lot of waste required to over cool the space. As well you can "short cycle" the cooling systems and really screw things up. By doing a contained cold isle, you get "some" control. As another post mentioned, it not perfect or 100% controlled because we do depend on the datacenter to do their job, but I think we are confident in the datacenter we are working with given 6 years of history.

G'Kar5
Gallente
Intaki Research and Manufacturing
Distorted Percepts
Posted - 2010.06.17 20:08:00 - [172]
 

Originally by: CCP Yokai

Dacil Arandur - What are the chances something like that could happen for this big move? (webcam for TQ move)

Answer - Sounds like fun, but have you seen us? Probably not the most exciting thing on cam and I couldn't handle the rejection if no one watched.



Please? With those 6748s, you'll surely have a free port or 2 for some webcams. You could stream TQ live 24/7. I'm sure some folks at CCP would use it. Maybe you can do something like http://www.visitphilly.com/museums-attractions/philadelphia/comcast-experience-video-wall-comcast-center/ at CCP headquarters, but rather a live feed of TQ. It would be fun, you know it!

OneBadMamaJama
Posted - 2010.06.17 20:08:00 - [173]
 

Originally by: Ticondrius
I've been here since the beginning, I've seen 2-3 total replacements now of the TQ cluster. I've heard the jokes about selling Oveur's car to buy the first RAMSAN. I've survived the Infiniband hype (which is apparenlty still not used...).

Push the Session Timer to 10 seconds (if not handle it more elegantly and transparent to the players) and redesign the session system as to permit multiple nodes to support a single star system, then I'll be impressed.

Software efficiency is ALWAYS better than throwing more hardware at the problem. I am not amused.


From a business standpoint, hardware is cheap, programmer time is expensive.

That being said, I find the "session changes" really annoying, especially when ships are destroyed. I'd like to see a refactoring to get rid of the in-space session system altogether, as this seems to be the cause of the most lag.

I'm also surprised that a game server that serves 40,000 players concurrently only has ~280 cores. That may explain why the NPC AI can hardly be called AI :)

Wouldn't have expected Cisco 7600s would've been used in a high-performance system like this either. Not that I know much about high-performance computing, but I used them in the networking classes I had to take, and figured they were crappy if that college could afford them :) I did like the console interface though.

Anyways, very cool of CCP to share this information about what goes on behind the scenes.

Mighty Birddy
Posted - 2010.06.17 22:30:00 - [174]
 

Your server specs mentioned Gigabit Ethernet as still being used. Where is it being used? (In respect to network topography and eve soller system mapping)

Demortes
Caldari
Oracle Phoenix
Posted - 2010.06.18 03:54:00 - [175]
 

Did I miss the answer to the OS question? What OS on the servers? Proprietary or not?

Kudos on the insight into EVE. I want a cluster like this to play all 60k accounts. :)

Olleybear
Minmatar
I R' Carebear
Posted - 2010.06.18 04:00:00 - [176]
 

Originally by: Commander Azrael
Originally by: Lord XSiV

... So much obnoxious stuff over many posts ...



I bet your an asbolute joy to work with.....




I must agree with you there about he must be an absolute joy...

To the "I'm smarter than thou art" poster,

It is truly amazing the level of arrogance in your posts sir. You took a wonderfull devblog about upgrades to tranquility and went so far over the top on the snob meter that I'm not sure if even superman can leap that in a single bound.

May the fleas of a thousand camels infest your armpits.

Tres Farmer
Gallente Federation Intelligence Service
Posted - 2010.06.18 04:00:00 - [177]
 

Originally by: CCP Yokai
Dacil Arandur - What are the chances something like that could happen for this big move? (webcam for TQ move)

Answer - Sounds like fun, but have you seen us? Probably not the most exciting thing on cam and I couldn't handle the rejection if no one watched.

If not live, even a timelapse of the whole thing would be fun to watch after.. Twisted Evil especially if it takes longer Wink
Just do it please!

Besides that - nice blog. Keep em coming!

Syberbolt8
Gallente
The Scope
Posted - 2010.06.18 04:03:00 - [178]
 

Edited by: Syberbolt8 on 18/06/2010 04:05:00
Originally by: Mighty Birddy
Your server specs mentioned Gigabit Ethernet as still being used. Where is it being used? (In respect to network topography and eve soller system mapping)


This doesn't really matter, The bottleneck is in the lack of processing power per node when its under heavy load. It hasn't been bandwidth for a long time, if ever.


Originally by: Demortes
Did I miss the answer to the OS question? What OS on the servers? Proprietary or not?

Kudos on the insight into EVE. I want a cluster like this to play all 60k accounts. :)


Last I heard they are running windows server 2008, though I could be wrong, I do know its windows server at least, and MSSQL for their database.

Zendoren
Aktaeon Industries
The Black Armada
Posted - 2010.06.18 07:12:00 - [179]
 

Edited by: Zendoren on 18/06/2010 07:38:35
Originally by: Demortes
Did I miss the answer to the OS question? What OS on the servers? Proprietary or not?

Kudos on the insight into EVE. I want a cluster like this to play all 60k accounts. :)


If you had that server you would have over 60k accounts ROFL =P

Originally by: Syberbolt8
Edited by: Syberbolt8 on 18/06/2010 04:05:00
Originally by: Mighty Birddy
Your server specs mentioned Gigabit Ethernet as still being used. Where is it being used? (In respect to network topography and eve soller system mapping)


This doesn't really matter, The bottleneck is in the lack of processing power per node when its under heavy load. It hasn't been bandwidth for a long time, if ever.


Originally by: Demortes
Did I miss the answer to the OS question? What OS on the servers? Proprietary or not?

Kudos on the insight into EVE. I want a cluster like this to play all 60k accounts. :)



Last I heard they are running windows server 2008, though I could be wrong, I do know its windows server at least, and MSSQL for their database.




As I remember from a Dev Post a while back, they are running Windows HPC Server 2008

Lord XSiV
Amarr
Kodar Innovations
Posted - 2010.06.18 09:05:00 - [180]
 

Originally by: Olleybear
Originally by: Commander Azrael

I bet your an asbolute joy to work with.....




I must agree with you there about he must be an absolute joy...

To the "I'm smarter than thou art" poster,

It is truly amazing the level of arrogance in your posts sir. You took a wonderfull devblog about upgrades to tranquility and went so far over the top on the snob meter that I'm not sure if even superman can leap that in a single bound.

May the fleas of a thousand camels infest your armpits.


It isn't like I care what you think. All I care (or cared about since I am now retired) was what the client thought. You see I was one of those people who got called in to clean up the mess that unqualified monkeys such as the previous posters complimenting Cisco junk would undoubtedly cause and ultimately affect an organizations bottom line. Or even worse, put lives at risk. No respected professional out there in the industry would ever take the risk of going with a Cisco solution when the implications of an outage were so great.

Originally by: Othran
Edited by: Othran on 17/06/2010 18:26:02
"While the routers we use are quite powerful (Cisco 7600's with the RSP720 route processors)"

Mmmm you sure you don't work for Cisco? Razz The last time I heard the term "powerful" and 7600 in the same sentence was a few years ago. Been a few years since I was in Triangle Park with the "folks".

Best of luck - there's always something physical that doesn't happen, hope yours is trivial.

Edit - as a sometime contractor for Cisco I can confirm that there is only one "Evil Empire" and its not Intel, MS or (recently) Apple Twisted Evil


Now this is funny and yet so true. Also brings back some good memories of RTP and 'playing' with some seriously interesting gear....

As for the 'Evil Empire' bit, don't worry, the largest IEC hasn't (nor will) drink the koolaid which will ensure that its potential dominance is kept in check. And you can always count on a good laugh when 'powerful', 'equipment' and 'Cisco' are mentioned in the same sentence without the word 'not'.

Anyhow the point still stands. The discussion is really about how much better CCP could be doing things if they elected to go with a better solution. The one that they are going to is far too costly, under performs (for proof just look at large fleet battles) which results in poor efficiency and a bad customer experience. By going with an efficient solution, over the long term CCP could save themselves considerable amounts of money which they then could use to up the pay scale to attract some of the better developers out there.



Pages: first : previous : ... 2 3 4 5 [6] 7 8 9 10 ... : last (11)

This thread is older than 90 days and has been locked due to inactivity.


 


The new forums are live

Please adjust your bookmarks to https://forums.eveonline.com

These forums are archived and read-only