Hooser: Narcos: datetimedate -+ interval integer timeperiod; #where timeperiod is second, minute, hour, day, month, year and some others. No plurals! egs., ‘SELECT NOW + INTERVAL 3 MONTH;’ ‘SELECT CURRENT_DATE – INTERVAL 1 DAY;’
Bredeson: Boleware: I don’t suppose you know if there’s a SQLAlchemy equivalent?
Pakonen: Hi, is there any good way to compare versions like x.x.x where x is any integer in an sql query?
Vernola: I cant use stored procedures
Diventura: Hi, I’m trying to create an index to optimise a select on a large table. my query is select foo, bar, baz etc from table where foo = ‘xxx’, bar=’yyy’, datetime_column = ‘2015-10-10 12:43:00’ order by primary_key_column limit 100000 – i paginate this through offsets. Is the best order for my index with the datetime column first?
Flatter: MarcBBxD: http://www.percona.com/files/presentations/percona-live/london-2011/PLUK2011-practical-mysql-indexing-guidelines.pdf
Aparo: What is wrong with this code?
Mullenaux: Http://pastie.org/private/x96e41a3euaqb8g4woi2ew
Constancio: It returns same line twice
Rappenecker: But there is only 1 line on ps_place table
Rumery: MrtAkdeniz, the join can multiply number of rows if one row in one table matches multiple rows from other table on join conditions
Evilsizor: Rumery, i didn’t get your point
Tabios: I mean, it can’t match multiple lines from other lines
Vincenzo: There is only 1 line from each table on condition
Lewallen: For example, there is only 1 ps_state which has id_state 1
Zdon: Rumery, thank you i found my problem, there were 2 different lanes which has different language id, that’s why there were 2 rows
Pruzansky: With 5.6.25-73.1 I have master-slave where on salve I have 300000 seconds behind master because of big DB which slowly decreasing. Looks logs are copied from Master to SLAVE and now what I see is that it’s only one CPU core in under load 90%. it’s single core load, any way to split this load to use all 24 cores on SLAVE server with 5.6 ?
Rumery: Eject_ck, by default replication is single-threaded, but 5.6 afaik supports parallel – http://dev.mysql.com/doc/refman/5.6/en/replication-options-slave.html#option_mysqld_slave-parallel-workers
Rumery: Eject_ck, but from the text is seems only updates to different databases can run at the same time
Rumery: Eject_ck, then 5.6 can probably use only one cpu to process the logs
Bostelmann: Means only way to improve performance is to scale-in
Tatsak: Hi guys, I was wondering how do slaves get master dumps for replication?
Hirt: Do the masters somehow feed the dumps via port 3306 to the slaves after it makes it?
Mcelderry: Hi. I have a bit of a situation with ibdata1 growing a bit too large a tiny little server, so was thinking of migrating the db to use myisam instead yes I know, not crash safe. If I just convert the tables to myisam, will I then be able to safely delete the ibdata1 and logfiles or will I need to keep them around for a while anyway?
Kujath: It will be safe if you disable innodb
Taormina: How about using file per table
Parilla: Boleware, using file per table does not really reduce the size it takes though? Just splits it up into multiple files
Rumery: Ether_Man, you should probably identify the reason why ibdata is growing – is some table placed in there growing significantly?
Rumery: Changing to myisam wont reduce it much too, same data take similar space
Christie: Rumery, considering it’s a zarafa mail server that’s using it, yes it’s growing.
Lechman: Does it use transactions or foreign keys?
Delaet: Oh? As I understand it, innodb uses ibdata for rollback, meaning it’s storing everything twice basically.
Ayling: Boleware, no idea. But it can use myisam as well as innodb so I would ***ume it does not.
Rumery: No, only active things – when the transaction is done those parts are reused
Rumery: But maybe you get some long running transactions or other problems in the app? but if it can work with myisam then it should run in autocommit and not be able to do that