| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10
Comments:
<0> myconid, per each person? .. ie. if subscriber list is 6000, 6000 queries?
<1> I mean, I can't sit the whole day in the mysql shell and watch if replication is working
<2> devnull: If your queries are only partially completing on the master, fix *that*, regardless of whether or not there is replication.
<1> mendel: uhm the master is working without problems?
<2> That's not what that error message you pasted said.
<2> It said a query partially completed on the master with error 1053.
<2> !perror 1053
<2> !error 1053
<2> argh, what's the syntax for that
<2> "Server shutdown in progress".
<1> well, that could be true
<2> If you want it to ignore that specific error, then...
<1> indeed
<2> on your slave, in my.cnf, slave_skip_errors=1053
<3> I'm working on a query, but I don't know how to do with my WHERE-clause in my MySQL query, this is my code: http://pastebin.com/573486
<3> I don't know how I can identify different rows in different tables
<1> means: skip those errors, or ignore them, or execute them but ignore the result?
<3> They have different ID cols, and some tables dont have ID cols at all
<3> :/
<2> devnull: Don't stop replication because of that error.
<4> UPDATE bolted_users SET usr_online=1 AND usr_login=current_timestamp() WHERE usr_id=1 <- errornous? seems like nothing get's updated
<0> myconid, I tested the speed of running seperate selects (looking up subscribers data) per user in the subscribers table, and it was extremely slow for a 5000+ list. Processing multiple rows seems to be much faster, though not as memory efficient. However I encountered a major problem
<1> mendel: oh, great thank you
<0> myconid, say I want to limit my search to 15 subscribers... if I do SELECT p.email,d.type,d.value FROM people p INNER JOIN data d using (pid) LIMIT 15; I could only get the 2 subscribers (say if they had 8 data each), and one w/ a partial set of data!
<5> Anyone here who has time?
<6> eml: s/and/,/
<4> DanGer, yeah, I figured afterwards, kind of stupid =)
<6> eml: and I think now() should work for usr_login
<4> DanGer, aren't they kinda the same ?:P
<6> think so, but now() far showrter :)
<4> yep =)
<4> well, it works
<7> mysql -user=horde -p***word=p*** < create.mysql.sql
<7> ERROR 1045: Access denied for user: 'ser=horde@localhost' (Using p***word: YES)
<0> cheloo, ya.. ?
<7> what does 'ser=horde@localhost' mean ?
<7> ser ?
<0> cheloo, sounds like you don't have a user setup named horde in the mysql server
<7> should be " user " and not ser
<7> yes i do
<7> have
<0> type: mysql -u horde -p
<0> see if you can login..
<1> uh another question about replication: I'm just replaying the binary log on the slaves, however I'm Seconds_Behind_Master: 1747 is that normal?
<7> Welcome to the MySQL monitor
<0> cheloo, you may have setup the user, but not given it a p***word..
<7> works fine
<7> bricecube i can login with the p***word i just setup
<0> cheloo, ahh ok.. do this: mysql -u horde -p <horde database name> < create.mysql.sql
<0> it'll ask you for p***wd.. then import the schema
<7> replace <horde database name> with what ?
<7> the database horde its empty
<0> cheloo, the database you want horde to use + the database the user horde has permissions on
<7> mysql -u horde -p horde < create.mysql.sql
<7> like this ?
<0> cheloo, the creaye.mysql.sql is there to populate your database..
<0> cheloo, if the database name is horde..
<7> yes
<8> is this room for mysql questions or mysql development?
<9> Hey you guys
<8> hello
<9> Anyone got time to help out? I'm trying to make a script to turn mixed tables (latin1/utf8) into pure utf-8
<9> This is how far I've come atm: http://pastebin.com/573486
<9> :-(
<9> PennStater got time to help me out?
<10> I could try, but I doubt I'd be able to ;o
<9> PennStater
<9> Anyone got time to help out? I'm trying to make a script to turn mixed tables (latin1/utf8) into pure utf-8
<10> heh, no clue, sorry :/
<7> bricecube ERROR 1044 at line 21: Access denied for user: 'horde@localhost' to database 'mysql'
<7> bricecube i get that if i run mysql -u horde -p horde < create.mysql.sql
<7> bricecube i set permission to grant CREATE,INSERT,DELETE,UPDATE,SELECT on hordedb.* to horde@localhost;
<7> so why i dont have access ?
<11> Hi, what sort of encryption does this looks like?
<9> Does what look like?
<11> 'AB6B9FFC2EAF54F1964C910E68094A15CF56C710105F0E2DF72D0691D5464F4B37E3684C298F97FB' ... it's too long for mysql p***word function, too long for SHA-1 and all caps (unlike md5)
<9> SHA-1 and MD5 aint encryption
<9> It could be any hash? :p
<11> but it's not mysql p***word() or md5 or sha1 right?
<9> nope
<9> but it could be any hash
<9> making a hash algoritm is fairly easy, someone could have made a hash on their own
<9> The chances of breaking it (if it's a hash), are fairly slimm I'd say :p
<11> I don't want to break it I want to create a new account into that DB and use the same encryption
<12> i'm desining a posts system, and i have the basics down, but i'm not sure how i should go about protecting them.. i had an awfull way before (seperating allowed usernames with a space) and am looking to change it.. any suggestions?
<13> How can i delete a database?
<12> rm -rf /
<14> hawking: drop database
<15> drop is a good option...
<13> th
<13> x
<15> hawking, sure..
<16> hola
<16> any ideas how this can be rewritten for performance ? http://www.sqlpaste.com/?entry_id=1019
<12> i'm desining a posts system, and i have the basics down, but i'm not sure how i should go about protecting them.. i had an awfull way before (seperating allowed usernames with a space) and am looking to change it.. any suggestions? what about having a table called.... allowed_users with columns: id, postid, userid ? seems a little.... to huge of a solution
<17> itrebal: That's a much better idea.
<12> thanks
<11> well Ill just try updating the p***sword field with an update query and the p***word() function and see if it works
<16> i guess there are no ideas ? :P
<18> Apachez: Er, why is SQL_NO_CACHE there if you want performance? :p
<19> how can i tell which row/s the result from mysql_result came from?
<19> erm mysql_query rather
<15> Dr1ft3d, use can check out a id so u will know from which row the reult came
<15> like select uid from ....
<0> cheloo, look at your database name,..
<0> Access denied for user: 'horde@localhost' to database 'mysql' ... what does that tell you?
<7> bricecube problem solved . thank you
<16> Detonator316: its example so i dont want the query cache to get hits when im developing the query
<16> its pretty bad when the query takes 45 seconds and next round just 10ms :P
<16> sure in production its all fine but not when you try to benchmark each method
<17> Apachez: The JOIN form might have better performance (in practice).
<16> how would a join form look like with my example ?
<16> what i try to accomplish is something like ( (select cols from t1) EXISTS (select cols from t2) EXISTS (select cols from t3) ) order by date limit 500;
<16> hmm
<16> EXISTS seems to be much faster than (col1, col2) IN (subquery.... ?
<17> Apachez: SELECT v1.f1, v1.f2 FROM () v1 JOIN (...) v2 ON (...) JOIN (...) v3 ON (...) ...;
<17> Apachez: Yes. EXISTS would help too.
<16> the query went like from 35 seconds down to 4 :)
<16> and all i changed was WHERE (col1, col2) IN (subquery..... into WHERE EXISTS (subquery....
<16> i guess they are equal or am i missing something by using EXISTS ?
<16> like any caveates ?
<17> The difference is EXISTS can stop after the first match in sub-processing.
<16> yeah im trying to run my two queries now, they should give me the same result
<16> if everything works my way :P
<17> If you have N-1 relationships, the first form will return duplicates in the result.
<17> Sorry. That's incorrect.
<16> hmpf... the results are not equal :/
<17> What's the difference?
<20> for a decimal range from 1.0 - 10.0 would i use DECIMAL(10,0) ?
<16> after the 5th row in the result the results are not equal
<17> Apachez: Different in what way?
<16> the rows returned
<16> could it be because the EXISTS method are using the outer query as "base" ?
<16> so with my limit 500 it will use the 500 rows from first query against the inner subquery and so on ?
<0> Is it possible to use a combination of group by & order targetting a specific column & an ***osiated columns value? ie. http://www.sqlpaste.com/?entry_id=1020
<17> Apachez: Make sure you account for ORDER differences that can result in different ordering in the output.
<16> hmm
<16> its seems that the AND (col1, col2) IN method is grouping the results based on col1 ?
<17> Apachez: Sorry. I forgot you had a limit. Yes. If your ordering is not explicit enough and you have relative duplicates, the result could be different.
<16> but bascially the EXISTS method "should" bring me ONLY id's which exist in all three tables right ?
<17> Apachez: The two are functionally identical without the LIMIT.
<17> Apachez: But you can't depend on order if you have relative duplicates.
<17> Apachez: If newsid is unique, order by that instead and check the results.
<16> yes but if i skip ORDER and LIMIT inside the subqueries and only use it in the most outer query then i should be safe ?
<16> yeah i need to do some more digging now
<17> Apachez: No.
<16> im comparing the results with my temptable method
<17> Apachez: The ordering in the subqueries is useless.
Return to
#mysql
or
Go to some related
logs:
yakuake crash xcomposite
ubuntu remove language packs
gimpshop .rpm
#debian
/etc/courier/shared/index
unexpected T_STRING APC
ubuntu ntp sync
rpm PUREFTP SUSE 10.1
fred merculis
change grub default edubuntu
|
|