#mysql
09 October 2007
Total 22 pages. You are browsing page 2/22.
First :: Prev :: [1] [2] [3] [4] [5] [...] :: Next :: Last
00:46 <****> i know how to tunnel
00:47 <****> its not workable, for other reasons ;)
00:47 <****> fizx: whats the error
00:47 <****> access denied
00:47 <****> the following works: mysql -u **** -h **** -P **** --password=**** --ssl --ssl-ca=SSL/cacert.pem --ssl-capath=SSL --ssl-cert=SSL/client-cert.pem --ssl-key=SSL/client-key.pem
00:48 <****> does it actualyl connect with ssl?
00:48 <****> \s
00:48 <****> SSL: Cipher in use is ...
00:49 <****> so yeah, i8 believe so
00:50 <****> i've ben following these instructions: http://dev.mysql.com/doc/refman/5.0/en/connector-j-reference-using-ssl.html
00:51 <****> so..
00:51 <****> keytool -import -alias mysqlServerCACert -file SSL/cacert.pem -keystore truststore
00:51 <****> keytool -genkey -keyalg rsa -alias mysqlClientCertificate -keystore keystore
00:52 <****> whats the exact error
00:52 <****> Recommendation for mysql HA? I want to do it with the fewest number of servers. SolidDB looks nice, anyone have any thoughts?
00:54 <****> i feel like i should be doing something different with keytool
00:54 <****> The driver encountered an error: java.sql.SQLException: Access denied for user 'casttv_db'@'74.95.2.89' (using password: YES)
00:55 <****> Free cookie if you help me http://paste.ubuntu-nl.org/40015/ Warning: mysql_fetch_assoc(): supplied argument is not a valid MySQL result resource in /var/www/freetest/action.php on line 20
00:55 <****> GO!
00:55 <****> okay.. hopefully this will be an intelligent question because I can't seem to find any answers on the web.
00:55 <****> free cokie if you do mysql_query(blabla) or die(mysql_error());
00:56 <****> if you have a large innodb table with an index that is not the primary key, and it has a million records added every week.. and you delete the old records every week (let's say as they become 2 months old)
00:56 <****> will that fragment or otherwise reduce the performance of your index over time?
00:56 <****> btw, you just posted your sql root password
00:57 <****> Sabrejack, the index shouldn't be hurt. The table might become fragmented. You can run OPTIMIZE TABLE tablename; to rearrange it nicely.
00:58 <****> hmm, cool deal
00:58 <****> Sabrejack, if performance is a big deal you could do something like create a table every day and pump your data into it. At the end of X days just drop the older tables so your index won't get super fragmented.
00:58 <****> hmm
00:59 <****> What pack suggests is faster in terms of the deletes, but probably a lot slower for selects, since you have to do lots of unions and probably filesorts over the derived table.
00:59 <****> well, luckily this is a session data table and the data is fully discardable
00:59 <****> Simetrical, i guess it depends on the contraints of your select. if you know you will be searching by date, you'd end up with a smaller select.
01:01 <****> and isn't really used in any joins etc.. so in the grand scheme of things I could probably automate new table creation
01:01 <****> pack, yes, it all depends. If you only usually do selects on the last few hours, you should do well enough.
01:01 <****> Also I'm used to the pg_vacum stuff that hoses your system, mysqls optimize may be less painful.
01:03 <****> oh, one more thing
01:03 <****> does replace into lead to more table fragmentation than update? I would expect it would since it deletes the old row, but I've never seen anyone actually say anything on the topic.
01:04 * Simetrical has no idea
01:04 <****> our entire data object base class uses replace into to save object data /facepalm
01:05 <****> I've never heard of it being a problem, but I'm not much of a MySQL expert, as things go.
01:05 <****> hmm
01:05 <****> InnoDB clusters by primary key, I've heard that.
01:05 <****> I don't think it cares how exactly you insert the data.
01:05 <****> yeah
01:06 <****> Never heard anyone say "it clusters by primary key except for when you REPLACE".
01:06 <****> I guess if you're replacing into with the same PK value it wouldn't fragment
01:06 <****> the update query sure does run like 10 times faster though
01:10 <****> ah crap
01:11 <****> please disregard. they changed the table type to myisam.
01:15 <****> Could someone please have a look here and tell me where I am making a mistake. It is only getting a "random" record in the first quarter or so of the questions table. http://pastebin.ca/730153
01:15 <****> I also think it is not very optimised.
01:15 <****> ERROR 1372 (HY000): Password hash should be a 41-digit hexadecimal number
01:15 <****> Trying to set password / create user with: GRANT USAGE ON *.* TO 'kself'@'%' IDENTIFIED BY PASSWORD 'literal-password';
01:15 <****> karsten: did you try to set the password directly?
01:15 <****> thumbs: ^^
01:16 <****> karsten: try password('literal-password')
01:16 <****> or passwd, I forget
01:16 <****> shamrock, your randomization assumes that id's are uniformly distributed. Google for "MySQL return random row" or somesuch and you'll see that this is a much-discussed problem.
01:16 <****> thumbs: ERROR 1064 (42000): You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near '('Kman') WITH GRANT OPTION' at line 1
01:17 <****> thumbs: Both ways.
01:17 <****> GRANT USAGE ON *.* TO 'kself'@'%' IDENTIFIED BY PASSWD('Kman') WITH GRANT OPTION;
01:17 <****> adduser
01:17 <****> hrrrm
01:17 <****> !man change password
01:17 <****> Sorry - I have no idea what function you're talking about! but try http://dev.mysql.com/change password
01:17 <****> !man grant
01:17 <****> see http://dev.mysql.com/doc/refman/5.0/en/grant.html
01:18 <****> Sim the UIDs are a linear sequence
01:18 <****> I'm working off my prior notes on doing this. How do I get mysql version w/in mysql?
01:18 <****> select version();
01:18 <****> karsten: read the grant man page
01:20 <****> hi, Im trying to do a fulltext search using MATCH AGAINSt, but the working of it is unclear to me, if I have an entry type text "blabla.txt" it returns this entry when I am querying for MATCH( column) AGAINST( '.txt*' IN BOOLEAN MODE), but not when I use '.txt' as the matching; anyone know why this is?
01:20 <****> Does the user have to exist before granting privs?
01:20 <****> thumbs: Reading.
01:21 <****> Doh!
01:21 <****> IDENTIFIED BY 'literal-password';
01:21 <****> No "PASSWORD" keyword.
01:23 * karsten wonder's what's buried in GRANT's tomb.
01:24 <****> Happy, happy, joy, joy.
01:27 <****> Um, how do I remove grants, mis-typed a database name.
01:27 <****> Deleting all grants where db == 'foo' would work.
01:28 <****> !man revoke
01:28 <****> see http://dev.mysql.com/doc/refman/5.0/en/revoke.html
01:28 <****> karsten: use that instead
01:31 <****> no fulltext search gurus about :)?
01:32 <****> thumbs: Thanks.
01:56 <****> Simetrical: or anyone, here is a working bit of PHP and mysql doing an efficient job of getting a random record and then getting the category name indexed by cid. http://pastebin.ca/730193 . But how to combine the two together so I only myke one DB trip. I realise its probably a join but I'll be damned if I can do it.
01:56 <****> hi
01:57 <****> shamrock, it's still not random if any rows have been deleted.
01:57 <****> if INSERT IGNORE has a key collision does it return the id of the record that it collided with?
01:57 <****> Simetrical: the randomness is good enough. its fast and there are no gaps.
01:58 <****> but can you educate me on how to make this a single query?
02:00 <****> What happens if you do this? SELECT questions.question, questions.answer, questions.cid, categories.name FROM categories JOIN questions ON categories.uid=questions.cid JOIN (SELECT ...) AS r2 USING (uid) ORDER BY categories.name LIMIT 1;
02:00 <****> Just sort of gluing them together. I'm too tired to think about the consequences properly.
02:01 <****> fuse_lt: wouldnt you already have the key if you tried to insert it?
02:01 <****> Of course it will require a filesort, I think.
02:01 <****> it is auto increment
02:01 <****> Hmm, will it? Not sure.
02:01 <****> if there is a collision, i want the primary key of the record it collided with
02:02 <****> fuse_lt, why would there be a collision if it's autoincrement?
02:02 <****> lets look at this
Total 22 pages. You are browsing page 2/22.
First :: Prev :: [1] [2] [3] [4] [5] [...] :: Next :: Last