| |
| |
| |
|
Page: 1 2 3
Comments:
<0> K_F- blogenator, blogster, mai-blog, pi-blog
<0> blawg
<0> sry if they're already taken
<0> K_F goes back to trying to be creative and figure names that might attract visitors
<0> ...('for his blogger')
<1> jessica blogalba
<1> I'll look twice
<0> hehe
<0> yeah
<0> blog-gen or blogen
<0> i think he's away
<1> blodgen - silent d (c:
<1> Ikow for a fact he aint here (c:
<0> ah
<0> he helped me with a sql problem earlier
<1> it's like 11 am for him... he'll be sleeping (c:
<1> he stays up late
<0> it's 12:40 am and i need to be up by 6
<0> i'm going to a filming
<0> of a star trek fan show
<0> so i don't have to be awake for it
<2> # /usr/local/bin/mysqladmin -u root -h myserver.mydomain.com p***word myp***word im getting an error
<2> saying access denied for user root@localhost
<3> trey
<3> try..
<3> mysql -h host -u user -p
<3> it'll ask you for the p***
<2> error 1130: Host' localhost.localdomain' is nt allowed to connect to this MYSQL server
<3> /:)
<3> what have you got there man?
<2> openbsd
<2> im able to ping the server
<2> i know mysql has started
<2> did a mysqladmin -u root p***word myp***
<2> after the installation
<3> hmm.. dunno
<4> hehe, just catched the prior conversation on the workstation :)
<4> the problem is that there is only one 400+ ops in the channel
<4> just in case X dissapears :)
<1> oh well... if/when it expires we'll just have to be nice to people to get 10 in
<1> I've got a pair... no sorry, I've got an eggy if you want (used to be a pair but I never got around to re-installing after a format)
<4> hehe
<4> remember, I'm a Linux-er
<1> Yeh, but i've notice you quit (c:
<4> K_F does... K_F- does not :)
<4> nor the server (which isn't on IRC in the first place)
<1> Are2Dee2 is hosted at work (c:
<1> oh so tired... well I've got a free p*** if I decide not to go to work tomorrow... I have a cold (which in turn leads to the body trying to fight it off, which leads to increased temperature, which in my case leads to chronic blinding headaches, which leads to pain killers that are capable of putting an elephant to sleep)
<4> hehe
<1> I over do everything, even a simple immune response (c:
<4> lol
<4> well, get better anyways :)
<1> you kidding? once the pain killers kick in, it's just a simple cold (c:
<1> btw that 2.5 hours summary process is now 15-20 mins long
<4> hehe
<1> between your suggestion and my re-write of the code
<4> nice
<1> it still takes 11 minutes for it to do a update set new=0 where new=1 and stamp < now - interval 4 day
<1> but that's ok, I'm trying to get a second machine to host the stuff older then 4 days then (at least for a while) it won't need to update 55+ million rows
<4> but would you need it anyways?
<4> at least the WHERE new=1 seems redundant
<1> it stops it from checking the timestamp on all the old stuff doesn't it?
<1> wouldn't it check the first condition before the second one?
<4> hmm
<4> probably
<1> it took 12 minutes as where stamp < now() -interval 4 day - but I wouldn't call it conclusive so it may be doing nothing
<1> once I get a second machine the problem will go away (until we're doing 12.5 million rows a day)
<4> you might have luck using a LIMIT clause
<4> what is the max that has ever been added in a day?
<1> I couldn't hesitate a guess. I work on averages per day - it takes one user to get on bittorrent to blow it out by 400 rows... and there's 600 users... so any number could do it
<4> lol
<4> how long does it take you to run SELECT COUNT(1) FROM table WHERE new = 1; ?
<1> my accounting is so incredibly complex I'm afraid... one day I'll draw it up - when I do I'll show you. You'll think it's disgusting but it's what I've had to do (c:
<1> lets find out shall we
<1> now we just sit back and wait
<1> SELECT COUNT(1) FROM table WHERE new = 1;
<1> 1 row in set (5.06 sec)
<1> not bad at all
<4> you can use that as a LIMIT
<1> so you're thinking I run the select count then...
<4> the update will never affect more rows
<1> update table set new=0 where stamp < now() - interval 4 day limit %count%
<1> might try that when I go to work just to see how much faster it is... cept I'll need to disable that crontab script...
<4> let me know how it turns out
<1> will do
<1> you know this'll prolly all go so much quicker if I wrote it as cpp (c:
<4> hehe
<4> what is it now?
<4> CFML?
<1> at the moment 1/4 is perl the other 3/4 is php (I know I know... but remmember I wanted to insure I could recycle the web based accounting code for the invoicing)
<4> I write virtually everything in PHP myself
<1> php isn't very suitable for p***ing 150 megs of data from one machine to another, then dealing with it (c:
<1> it was 43% faster in perl
<1> it'll probably be that again in cpp
<4> nah, you write a cpp extension for PHP
<4> :)
<1> that's how I was going to start handling encryption
<1> up until now I'd been using eaccelerator and encoding the encryption.php file.
<1> but yes that's a brilliant idea
<4> anyways, I'm off
<4> good luck with your cold and your code
<4> have a nice day
<1> thought I could smell soemthing
<1> take it easy mate (c:
<1> doh
<1> I just found a problem with using the count() trick
<4> oh?
<4> yeah, I came back
<1> I have to make sure no-other processes are adding to the table at the same time
<4> add a safety margin
<1> fetch runs every 15 minutes...or as often as it can
<1> could re-invent the cron so to speek
<1> have it run on a queue based system so that the three primary scripts can't conflict and run all at once (at 1:30 am)
<4> how many records did the COUNT(1) return?
<4> can simply do something like SELECT COUNT(1) * 1.05 , giving you a 5% margin
<1> | 5746131 | <- first run
<1> | 5767242 |
<1> second one
<4> a 1% margin is enough then
<1> I think * it by 1.05 but still stop all the other scripts from running at 1:30
<1> if the processing at 1:30 breaks for any reason we don't send out invoices... that means I have to do it manually (c:
<4> better be sure then :)
<4> but a 5% margin is too much for that data set
<4> go for 2-3 max
<1> spoze if I'm stopping all other work it's ok to go for 1%
<1> I already had to stop summary and processing from going off at the same time or peoples summaries would do weird things (C: (as half the data was marked as old while it was being run)
<4> hehe
<5> 1@find mysql
<6> hi
<6> please how to connect to mysql ?
<6> via net
<6> i'have mysql server name
<6> and user name
<6> and p***word
<6> but i'can't connect
<6> pls help me
<4> accepted hostname
<4> and spell out the words in here
<6> what?
<4> "<6> pls help me"
<4> spell out the words in here.
<6> because i'm bloqued
<6> ok
<4> MySQL operates with a whitelist, not a blacklist
<4> you will find all the required information in the manual.
<6> i'have all information server name, user name and p***word
<6> but it can't connect
Return to
#mysql
or
Go to some related
logs:
#c
#windows
#AllNiteCafe
#MissKitten
megadeth disconnect.mp3
#c++
#c++
#linux
#linuxhelp
fusionroom
|
|