| Pages: 1 2 [3] :: one page |
|
|
| Author |
Thread Statistics | Show CCP posts - 6 post(s) |
Ford Mersombre
Gallente
Federal Navy Academy
   |
Posted - 2010.07.28 12:39:00 -
[61]
Another DB error? o.O
I think it's a conspiracy. I'd bet money the devs are really out stealing minerals from POS's and sticking bananas up Titian tail pipes.
|
slave redeemer
Minmatar
|
Posted - 2010.07.28 12:39:00 -
[62]
I`m in......I`m ****in in I tell ya |
Takseen
|
Posted - 2010.07.28 12:42:00 -
[63]
Just wanted to say that I appreciate these quick status updates, and I hope you keep making them.
|
HottyChick
|
Posted - 2010.07.28 12:44:00 -
[64]
I was gonna reboot TQ on time till I got hi** ;=])
|
Tanjia Guileless
|
Posted - 2010.07.28 12:48:00 -
[65]
Another MS SQL Server hits the dust.
|
NoChance Lance
|
Posted - 2010.07.28 12:52:00 -
[66]
 Originally by: Ford Mersombre
Another DB error? o.O
I think it's a conspiracy. I'd bet money the devs are really out stealing minerals from POS's and sticking bananas up Titian tail pipes.
I'm not massively sure Titian would have been too fond of that.
_______________________________________________
I'm not afraid of anything or anyone, unless they are wearing a mask. Diego Maradona 2010 |
|
CCP Yokai
|
Posted - 2010.07.28 12:56:00 -
[67]
What happened today and last week was that the query optimizer came to the wrong conclusions about that would be the best plan of action for loading solar systems.
We were getting a bad query plan and index fragmentation does not change the results of the query compilation, but somewhere we hit a so called 'tipping point', a point where the query optimizer suddenly changes its mind on what is best to do.
This often happens due to the number of rows or pages in a table since that is what one part of the decision is based on. This also happens if there is a certain number of rows within the range the query has to scan or seek on and that's where statistics updating helps.
What we did was to override the decisions of the query optimizer and told it to use a certain way of comparing two tables.
|
|
Makko Gray
Gallente
Nexus Aerospace Corporation
The Volition Cult
|
Posted - 2010.07.28 13:00:00 -
[68]
Edited by: Makko Gray on 28/07/2010 13:01:33
Originally by: CCP Yokai
What happened today and last week was that the query optimizer came to the wrong conclusions about what would be the best plan of action for loading solar systems.
We were getting a bad query plan and index fragmentation does not change the results of the query compilation, but somewhere we hit a so called 'tipping point', a point where the query optimizer suddenly changes its mind on what is best to do.
This often happens due to the number of rows or pages in a table since that is what one part of the decision is based on. This also happens if there is a certain number of rows within the range the query has to scan or seek on and that's where statistics updating helps.
What we did was to override the decisions of the query optimizer and told it to use a certain way of comparing two tables.
Is that a subtle way of saying 'something broke but we ran the database engine tuning advisor in sql management studio and it seems to be working now'?
Yeah, I do that 
Edit: Worth going over the execution plans manually as well, its suggestions can't always be trusted.
|
Ressiv
Cooperative Freelance Navigators Association
|
Posted - 2010.07.28 13:04:00 -
[69]
Originally by: CCP Yokai
...
Aah .. the wonders of MS SQL tricks & gadgets havent ceased to disapoint people yet I see ?
==========================
Nothing is true, everything is permitted.
========================== |
Teibor
|
Posted - 2010.07.28 13:04:00 -
[70]
Hmm, more like a subtle way of saying 'we are using something which isnt up to the job so occasionally we need to hit it with a big stick to get it to work'.
|
|
|
T'ess Emalee
|
Posted - 2010.07.28 13:07:00 -
[71]
Originally by: CCP Yokai
What happened today and last week was that the query optimizer came to the wrong conclusions about what would be the best plan of action for loading solar systems.
We were getting a bad query plan and index fragmentation does not change the results of the query compilation, but somewhere we hit a so called 'tipping point', a point where the query optimizer suddenly changes its mind on what is best to do.
This often happens due to the number of rows or pages in a table since that is what one part of the decision is based on. This also happens if there is a certain number of rows within the range the query has to scan or seek on and that's where statistics updating helps.
What we did was to override the decisions of the query optimizer and told it to use a certain way of comparing two tables.
Why didn't you fix it last week?
If you know what the problem is fix it.
|
Ressiv
Cooperative Freelance Navigators Association
|
Posted - 2010.07.28 13:09:00 -
[72]
I wonder hom much performance could be gained by just dropping the MS servers and dabatase, and replace it with a Unix/PostgreSQL environment.
Yeah, it would probably mean that a couple of monkeys would be out of a job, and skilled people needed in return, but that never hurt anyone. (No, not even the monkeys in the long run!)
But meh, wont happen..
==========================
Nothing is true, everything is permitted.
========================== |
SHkjell
|
Posted - 2010.07.28 13:34:00 -
[73]
When i try to log in , i get a black screen with a message "entering station" and a status bar that is 100% filled , but then nothing happens , i have waited for 7 minutes now. :(
|
Ted Warrior
|
Posted - 2010.07.28 13:39:00 -
[74]
Originally by: SHkjell
When i try to log in , i get a black screen with a message "entering station" and a status bar that is 100% filled , but then nothing happens , i have waited for 7 minutes now. :(
Same for me :(
|
RifterDrifter
They Found Oil On Your Anus
|
Posted - 2010.07.28 14:04:00 -
[75]
Originally by: CaptCole
Why did you move the cluster to London? O I forgot better performance????
Ever since you have done this you have had a lot of trouble keeping your customers happy.
I ****ING KNEW IT! It's always the Island Monkeys! Nuke England!
Originally by: CCP Yokai
query optimizer came to the wrong conclusions about what would be the best plan of action for loading solar systems.
My client seems to have the same problems loading solar systems every time my fleet jumps into an enemy fleet. Can you peeps help me with that?
_______________________________________________
|
Sublunar
Caldari
Capitalism and Freedom
|
Posted - 2010.07.28 14:58:00 -
[76]
Edited by: Sublunar on 28/07/2010 15:00:28
Thank you CCP for keeping us updated.
I really appreciate that.
@Makko Gray
check for t-sql hints at
http://msdn.microsoft.com/en-us/library/ms187713.aspx
|
Sordeo Regius
Blueprint Haus
Get Off My Lawn
|
Posted - 2010.07.28 15:07:00 -
[77]
Feed the hamsters darn it!
|
YarrMama
|
Posted - 2010.07.28 15:26:00 -
[78]
Originally by: CCP Yokai
What happened today and last week was that the query optimizer came to the wrong conclusions about what would be the best plan of action for loading solar systems.
We were getting a bad query plan and index fragmentation does not change the results of the query compilation, but somewhere we hit a so called 'tipping point', a point where the query optimizer suddenly changes its mind on what is best to do.
This often happens due to the number of rows or pages in a table since that is what one part of the decision is based on. This also happens if there is a certain number of rows within the range the query has to scan or seek on and that's where statistics updating helps.
What we did was to override the decisions of the query optimizer and told it to use a certain way of comparing two tables.
Well I see what your problem is - You've got the darn Flux Capicitor in backwards. Back this thing up to 88mph and you should be able to roll back to when this game was semi-stable.
I think we need somemore free skill points to help our optimizer deal with all the extended downtimes.
|
|
CCP Navigator
C C P
C C P Alliance
|
Posted - 2010.07.28 16:49:00 -
[79]
Now that TQ has been up for 3 hours with no issues this thread will be locked.
Navigator
Senior Community Representative
CCP Hf, EVE Online
|
|
|
|
| |
|
| Pages: 1 2 [3] :: one page |
| First page | Previous page | Next page | Last page |