| Pages: 1 2 [3] 4 :: one page |
| Author |
Thread Statistics | Show CCP posts - 6 post(s) |

skandalf
|
Posted - 2006.08.29 05:15:00 -
[61]
lol....
the server is up its just that your character is stuck
|

Eis Bachs
Karnival of Death Squad
|
Posted - 2006.08.29 05:16:00 -
[62]
Originally by: Lysithea Edited by: Lysithea on 29/08/2006 05:12:35 Oh come on... don't say the server is back up then give me this lame message when i try to log in:
"Your character is located within (Otitoh), which is currently loading. Please try again in a moment."
Either bring the cluster up or don't. This half-assed attempt is bad. This is like jumping into my car in the morning to go to work and being told "two wheels don't work right now". Not much help there.
Then my other character tries to log in and is stuck at the login screen.
Please, just take the extra 10 minutes and get the whole cluster up. Stop the mediocrity.
I agree, though clearing that abortion of a cache file called machonet should help. CCP, please fix the state recognition of the client. kthnxbai
|

Caryna
Caldari
|
Posted - 2006.08.29 05:16:00 -
[63]
Originally by: ThunderGodThor
Originally by: aggiedog Could we fly the Chinese server over to London and swap out the servers?
.... Its the DB and all its relaited things. The size of eve is the problem and untill the migration all the way over to the 64bit coding it wont be really fixed...
right, MySQL can't handle such large DB's very well. think about migrate to an other DB software, like ORACLE 
|

Lazuran
|
Posted - 2006.08.29 05:16:00 -
[64]
Upgrade time: from 64GB to http://www.superssd.com/products/tera-ramsan/
"The whole of NYC is not 1.0. Some back alley in the Bronx is deep 0.0, while right outside NYPD headquarters is 1.0." -- Slaaght Bana |

Sirial Soulfly
|
Posted - 2006.08.29 05:18:00 -
[65]
Does Valar ever sleep ?
|

hendo001
Caldari Quantum Tech Mining
|
Posted - 2006.08.29 05:18:00 -
[66]
Edited by: hendo001 on 29/08/2006 05:19:15 Thx for the update valar, did they get u out of bed to sort it??
Hope the new Yarrrdware arrives soonÖ
edit: spellings n stuff
QTM is currently recruiting miners convo me or send an evemail in game Visit our forums Linkage
Link removed - Wrangler |

hendo001
Caldari Quantum Tech Mining
|
Posted - 2006.08.29 05:20:00 -
[67]
Originally by: Sirial Soulfly Does Valar ever sleep ?
yeah in the ccp office above the servers as its nice n warm there   
QTM is currently recruiting miners convo me or send an evemail in game Visit our forums Linkage
Link removed - Wrangler |
|

Valar

|
Posted - 2006.08.29 05:22:00 -
[68]
Originally by: hendo001 Edited by: hendo001 on 29/08/2006 05:19:15 Thx for the update valar, did they get u out of bed to sort it??
Hope the new Yarrrdware arrives soonÖ
edit: spellings n stuff
Yeah, they got me out of bed to sort it. And now I'm going back to bed. ------ Valar Database admin - Server operations team CCP Games How to write a good bugreport |
|
|

Valar

|
Posted - 2006.08.29 05:24:00 -
[69]
Note that, as after any startup, solarsystems will take some time to load, but especially now since the database failed over when the transaction log filled up, so its cache and buffers are all empty, so all data has to be read from its disks. ------ Valar Database admin - Server operations team CCP Games How to write a good bugreport |
|

Lazuran
|
Posted - 2006.08.29 05:26:00 -
[70]
can't sell tritanium at the moment:
Screenshot
"The whole of NYC is not 1.0. Some back alley in the Bronx is deep 0.0, while right outside NYPD headquarters is 1.0." -- Slaaght Bana |

hendo001
Caldari Quantum Tech Mining
|
Posted - 2006.08.29 05:27:00 -
[71]
Edited by: hendo001 on 29/08/2006 05:27:58
Originally by: Valar
Originally by: hendo001 Edited by: hendo001 on 29/08/2006 05:19:15 Thx for the update valar, did they get u out of bed to sort it??
Hope the new Yarrrdware arrives soonÖ
edit: spellings n stuff
Yeah, they got me out of bed to sort it. And now I'm going back to bed.
Hope u get paid a call out fee m8
QTM is currently recruiting miners convo me or send an evemail in game Visit our forums Linkage
Link removed - Wrangler |

SonOTassadar
The Dead Parrot Shoppe Inc.
|
Posted - 2006.08.29 05:30:00 -
[72]
Originally by: Valar Note that, as after any startup, solarsystems will take some time to load, but especially now since the database failed over when the transaction log filled up, so its cache and buffers are all empty, so all data has to be read from its disks.
Valar: I think a sticky needs to go up addressing the issue at hand. I've seen more and more threads lately complaining about how the server is up but they can't get into the system they're in (it's still loading, as we've all seen). While I don't mind this, a number of people here feel that everything must work at exactly the moment you say it does. 
Furthermore, adding to the sticky that new hardware is on the way may help ease the current influx of threads complaining about the server. ___________________________ |

hendo001
Caldari Quantum Tech Mining
|
Posted - 2006.08.29 05:35:00 -
[73]
Originally by: SonOTassadar
Originally by: Valar Note that, as after any startup, solarsystems will take some time to load, but especially now since the database failed over when the transaction log filled up, so its cache and buffers are all empty, so all data has to be read from its disks.
Valar: I think a sticky needs to go up addressing the issue at hand. I've seen more and more threads lately complaining about how the server is up but they can't get into the system they're in (it's still loading, as we've all seen). While I don't mind this, a number of people here feel that everything must work at exactly the moment you say it does. 
Furthermore, adding to the sticky that new hardware is on the way may help ease the current influx of threads complaining about the server.
I have to agree with this, its an excellent point.
Also when is the new yarrrrdware due?
QTM is currently recruiting miners convo me or send an evemail in game Visit our forums Linkage
Link removed - Wrangler |

MysticNZ
Solstice Systems Development Concourse
|
Posted - 2006.08.29 07:06:00 -
[74]
Originally by: Lazuran Edited by: Lazuran on 29/08/2006 05:40:08 Edited by: Lazuran on 29/08/2006 05:29:25 can't sell tritanium at the moment:
Screenshot
Edit: can't even move it back to my cargo ... same message.
Edit: relogging fixed it
Bug reported this two days ago. You can make it happen quite easily. -=====-
|

Shittake
RONA Deepspace Rule of Three
|
Posted - 2006.08.29 07:12:00 -
[75]
Originally by: Valar Due to the fast growth of EVE, the database grows at a much faster rate then before. This means that we need to do more maintenance downtimes and can lead to incidents like what happened now. With the biggest tables and the transaction log on the RAMSAN, we get dangerously low on diskspace there between extended maintenance downtimes, and tonight everything went boom when the transaction log filled up while a full backup was running on the database.
I've put in temporary measures to prevent this from happening again... well, prevent it from crashing the cluster because of this and we have more hardware on the way.
I'm really sorry for any inconveniance this downtime caused.
While I am just a small-potatoe server tech in the real world, one of things I regulary do for my customers is capacity planning. It's not a very tough thing to do, and if its done correctly you NEVER run out of disk space.
You are either NOT listening to your IT support folks on their reccomendations OR they are failing at their jobs - either way your customer base is suffering from CCP's lack of capacity planning.
This is not rocket science. GET MORE RAMSAN SPACE NOW!!!!
Am I the only IT support guy who plays this game who sees this?
|

Celiss
|
Posted - 2006.08.29 07:56:00 -
[76]
What bothers me more than the failure for good capacity is what seems like the lack of monitoring of a known recurring condition.
It seems that running out of space due to growing transaction logs has brought things down several times now. Seems like it'd be nice if they got pages/SMSes when they're down to their last 5% or some other critical level, so they can go fix the problem before it takes the form of a messy crash. Even if the server still had to be restarted, a "server will restart in 15 minutes" would be a big improvement over the outright crashes. And help reduce the massive number of petitions that flow into customer service everytime the server crashes.
|

DukeJoost1
The Last Solution Inc
|
Posted - 2006.08.29 07:59:00 -
[77]
Originally by: ****take
This is not rocket science. GET MORE RAMSAN SPACE NOW!!!!
Am I the only IT support guy who plays this game who sees this?
No, I see it too. And seen at many of my customers. Doing storage management.... Capacity planning indeed.
Fly safe.
 |

Matthew
Caldari BloodStar Technologies
|
Posted - 2006.08.29 08:00:00 -
[78]
Originally by: ****take While I am just a small-potatoe server tech in the real world, one of things I regulary do for my customers is capacity planning. It's not a very tough thing to do, and if its done correctly you NEVER run out of disk space.
Capacity planning is always based on estimates of what will happen. Estimates are never exact, no-one has the ability to see the future. If growth outstrips estimates by a significant amount, even generous safety margins can be exceeded.
Of course, if that starts happening in a small-potato world, it's usually quick and easy to grab some more drives and bung them in the disk array. Standard drives, controllers etc are always in stock and available on short notice. Strangely enough, RAMSAN's are not available for next-day delivery. There won't be stocks sitting around the place because it's such a high value, low volume item. I'd be surprised if the company even started making it until they had the order in. Combine that with shipping it from the US to Iceland for testing, then Iceland to London for installation, and you're looking at a lead-time measured in months.
So they may well have ordered another RAMSAN back before this even became a player-visible problem, and are now at the mercy of time factors that they have no control over. ------- There is no magic Wand of Fixing, and it is not powered by forum whines. |

Miss Overlord
|
Posted - 2006.08.29 09:08:00 -
[79]
yes capacity seems to be the thing (removal of 12 month old inactive accounts is a good start)
|

Helison
Gallente Times of Ancar R i s e
|
Posted - 2006.08.29 09:40:00 -
[80]
Originally by: Miss Overlord yes capacity seems to be the thing (removal of 12 month old inactive accounts is a good start)
Removal of all BM`s might be a better start.
|

PhamNuwen
Caldari Bungee Jumper
|
Posted - 2006.08.29 10:24:00 -
[81]
Edited by: PhamNuwen on 29/08/2006 10:26:21
Originally by: Valar Due to the fast growth of EVE, the database grows at a much faster rate then before. : ...and tonight everything went boom when the transaction log filled up while a full backup was running on the database.
I've put in temporary measures to prevent this from happening again... well, prevent it from crashing the cluster because of this and we have more hardware on the way. ...
Script: if (LowOnDiskSpace) then CancelBackup; //remarked DBA: we did never need the backups...
__
still wating for infos about patches and the arrival of new content  |

Miss Overlord
Gallente Garoun Investment Bank
|
Posted - 2006.08.29 11:25:00 -
[82]
Originally by: Helison
Originally by: Miss Overlord yes capacity seems to be the thing (removal of 12 month old inactive accounts is a good start)
Removal of all BM`s might be a better start.
i support both ideas if not entire BM remove a limit of 5000 or 10000 per character.
|

storm2k5
|
Posted - 2006.08.29 11:25:00 -
[83]
Originally by: ****take
Originally by: Valar Due to the fast growth of EVE, the database grows at a much faster rate then before. This means that we need to do more maintenance downtimes and can lead to incidents like what happened now. With the biggest tables and the transaction log on the RAMSAN, we get dangerously low on diskspace there between extended maintenance downtimes, and tonight everything went boom when the transaction log filled up while a full backup was running on the database.
I've put in temporary measures to prevent this from happening again... well, prevent it from crashing the cluster because of this and we have more hardware on the way.
I'm really sorry for any inconveniance this downtime caused.
While I am just a small-potatoe server tech in the real world, one of things I regulary do for my customers is capacity planning. It's not a very tough thing to do, and if its done correctly you NEVER run out of disk space.
You are either NOT listening to your IT support folks on their reccomendations OR they are failing at their jobs - either way your customer base is suffering from CCP's lack of capacity planning.
This is not rocket science. GET MORE RAMSAN SPACE NOW!!!!
Am I the only IT support guy who plays this game who sees this?
You know, non-potatoe-ramsan-disks aren't exactly cheap.
|

Tetra gramaton
Gallente
|
Posted - 2006.08.29 12:21:00 -
[84]
Arghhhh this is so frutrating to be an aussie gamer in eve. It goes down EVERY night at 9pm and is suppose to be back up at 10pm... Wow 50% of the time after 10 I cant get back in. I give up after 10.30pm and figure it not worth trying any more :(
|

Riho
Mercenary Forces
|
Posted - 2006.08.29 12:26:00 -
[85]
Originally by: Valar Note that, as after any startup, solarsystems will take some time to load, but especially now since the database failed over when the transaction log filled up, so its cache and buffers are all empty, so all data has to be read from its disks.
makes perfect sense :D
ty for the info :)
|

w0rmy
Intensive CareBearz
|
Posted - 2006.08.29 12:29:00 -
[86]
Originally by: storm2k5
You know, non-potatoe-ramsan-disks aren't exactly cheap.
Especially when theyre all being bought for the china cluster!
Originally by: Avon
Originally by: Dark Shikari
What single item is larger than a jetcan?
My ego?
|

storm2k5
|
Posted - 2006.08.29 12:42:00 -
[87]
Originally by: Jim McGregor
Current status: Market item listings not responding. Undock works, mailbox works, wallet works.. everything works except market.
Give it a few minutes m8, always the same after DT
|

Jim McGregor
Caldari
|
Posted - 2006.08.29 12:43:00 -
[88]
Edited by: Jim McGregor on 29/08/2006 12:44:09
Originally by: storm2k5
Originally by: Jim McGregor
Current status: Market item listings not responding. Undock works, mailbox works, wallet works.. everything works except market.
Give it a few minutes m8, always the same after DT
I think it worked pretty well there for a few days actually... maybe it was the queues helping things. --- Eve Wiki | Eve Tribune | Eve Pirate |

Pookie McPook
Nova Synergy PLC
|
Posted - 2006.08.29 13:57:00 -
[89]
Originally by: Valar Due to the fast growth of EVE, the database grows at a much faster rate then before. This means that we need to do more maintenance downtimes and can lead to incidents like what happened now. With the biggest tables and the transaction log on the RAMSAN, we get dangerously low on diskspace there between extended maintenance downtimes, and tonight everything went boom when the transaction log filled up while a full backup was running on the database.
I've put in temporary measures to prevent this from happening again... well, prevent it from crashing the cluster because of this and we have more hardware on the way.
I'm really sorry for any inconveniance this downtime caused.
Maybe CCP should think about upgrading their 286's 
Seriously though....we gamers are constantly being told that we need to upgrade to higher and higher grade spec machines in order to run games...perhaps the gaming companies should consider this when selecting their servers? We ARE paying our subscription for a reason you know.... not just to be divided out by way of dividends, share options or corporate hospitality in Bermuda.
Perhaps the grail of a shardless universe is not tenable with current technology after all. 
|

storm2k5
|
Posted - 2006.08.29 14:09:00 -
[90]
Originally by: Pookie McPook
Originally by: Valar Due to the fast growth of EVE, the database grows at a much faster rate then before. This means that we need to do more maintenance downtimes and can lead to incidents like what happened now. With the biggest tables and the transaction log on the RAMSAN, we get dangerously low on diskspace there between extended maintenance downtimes, and tonight everything went boom when the transaction log filled up while a full backup was running on the database.
I've put in temporary measures to prevent this from happening again... well, prevent it from crashing the cluster because of this and we have more hardware on the way.
I'm really sorry for any inconveniance this downtime caused.
Maybe CCP should think about upgrading their 286's 
Seriously though....we gamers are constantly being told that we need to upgrade to higher and higher grade spec machines in order to run games...perhaps the gaming companies should consider this when selecting their servers? We ARE paying our subscription for a reason you know.... not just to be divided out by way of dividends, share options or corporate hospitality in Bermuda.
Perhaps the grail of a shardless universe is not tenable with current technology after all. 
God, they are not stupid. You really think they didn't consider to get more space for the transaction logs? 1GB of ramsan costs roughly 2000$, that's without implementation. They currently have 64GB ramsan-400 if I'm not mistaken, and that baby costs 130'000$
|
| |
|
| Pages: 1 2 [3] 4 :: one page |
| First page | Previous page | Next page | Last page |