 | |
No autocommit in JDBC driver? extra speed?
mailing.database.mysql-java mailing.database.mysql-java
 |
| | Thread Tools | Display Modes |
#1
 
| |||
| |||
|
No autocommit in JDBC driver? extra speed? -
05-28-2004
, 05:59 PM
#2
| |||
| |||
|
|
I enabled profileSql=true and noticed: Query "SET autocommit=1" execution time: 33 Query "SET autocommit=1" execution time: 36 Query "SELECT * FROM FEED WHERE ID = '1'" execution time: 37 result set fetch time: 0 What's with all the autocommits? Isn't this redundant on MyISAM tables?. Is there any way to disable this via the connection URL? Doesn't look like it: http://dev.mysql.com/doc/connector/j...html#id2801179 Kevin FYI... this was because we were using connection pooling. Our |
this to the server.
The attached patch makes Connection.java smarter and if it's already
been sent and it's the same value we don't resend it. Should be simple
and clean...
Thoughts on this guy?
--
Please reply using PGP.
http://peerfear.org/pubkey.asc
NewsMonster - http://www.newsmonster.org/
Kevin A. Burton, Location - San Francisco, CA, Cell - 415.595.9965
AIM/YIM - sfburtonator, Web - http://peerfear.org/
GPG fingerprint: 5FB2 F3E2 760E 70A8 6174 D393 E84D 8D04 99F1 4412
IRC - freenode.net #infoanarchy | #p2p-hackers | #newsmonster
--------------000802080305060005050409
Content-Type: text/plain;
name="Connection.java.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
filename="Connection.java.patch"
46d45
<
203a203,205
Quote:
|
/** Has autocommit already been set for this connection? */ private boolean hasAutoCommit = false; 509a512,517 //prevent duplicate autocommit settings which will just slow our //connection to the server if ( hasAutoCommit && this.autoCommit == autoCommit ) return; 515d522 |
548a556,557
Quote:
|
hasAutoCommit = true; 2724d2732 |
3129d3136
<
--------------000802080305060005050409
Content-Type: text/plain; charset=us-ascii
--
MySQL Java Mailing List
For list archives: http://lists.mysql.com/java
To unsubscribe: http://lists.mysql.com/java?unsub=my...ie.nctu.edu.tw
--------------000802080305060005050409--
#3
| |||
| |||
|
|
Kevin Burton wrote: I enabled profileSql=true and noticed: Query "SET autocommit=1" execution time: 33 Query "SET autocommit=1" execution time: 36 Query "SELECT * FROM FEED WHERE ID = '1'" execution time: 37 result set fetch time: 0 What's with all the autocommits? Isn't this redundant on MyISAM tables?. Is there any way to disable this via the connection URL? Doesn't look like it: http://dev.mysql.com/doc/connector/j...html#id2801179 Kevin FYI... this was because we were using connection pooling. Our persistence layer kept calling setAutoCommit( true ) which kept sending this to the server. The attached patch makes Connection.java smarter and if it's already been sent and it's the same value we don't resend it. Should be simple and clean... Thoughts on this guy? ------------------------------------------------------------------------ 46d45 203a203,205 /** Has autocommit already been set for this connection? */ private boolean hasAutoCommit = false; 509a512,517 //prevent duplicate autocommit settings which will just slow our //connection to the server if ( hasAutoCommit && this.autoCommit == autoCommit ) return; 515d522 548a556,557 hasAutoCommit = true; 2724d2732 3129d3136 |
It might be interesting, but at first glance it doesn't appear that this
patch 'knows' if auto commit has been set _outside_ of the JDBC
calls...the only 'safe' way to know this is to issue it explcitly.
This also seems like a premature optimization, are you _really_ sure the
sub-ms call to setAutoCommit() is causing your application to be slower,
compared to query optimization, caching, etc that you could do that
generally have orders-of-magnitude performance improvement possibilities?
-Mark
- --
Mr. Mark Matthews
MySQL AB, Software Development Manager, J2EE and Windows Platforms
Office: +1 708 332 0507
www.mysql.com
MySQL Guide to Lower TCO
http://www.mysql.com/it-resources/white-papers/tco.php
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (MingW32)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFAu5TJtvXNTca6JD8RAtvpAJ9nqoUN3467rBuyfj0UO8 3N4w2pjgCeMmMg
bNHNqUgbxbjsoYLpFE95RNU=
=biCL
-----END PGP SIGNATURE-----
--
MySQL Java Mailing List
For list archives: http://lists.mysql.com/java
To unsubscribe: http://lists.mysql.com/java?unsub=my...ie.nctu.edu.tw
#4
| |||
| |||
|
|
Kevin, It might be interesting, but at first glance it doesn't appear that this patch 'knows' if auto commit has been set _outside_ of the JDBC calls... Oh... is there a difference if the com.mysql.* classes have set |
multiple true or false are coalesced.
From looking at the code a call from com.mysql.* vs another 3rd party
wouldn't change anything even if they were coalesced.
Quote:
|
the only 'safe' way to know this is to issue it explcitly. It is issued... but we just remove duplicate SET AUTOCOMMIT=true... |
Quote:
|
This also seems like a premature optimization, are you _really_ sure the sub-ms call to setAutoCommit() is causing your application to be slower, compared to query optimization, caching, etc that you could do that generally have orders-of-magnitude performance improvement possibilities? Oh no... let me make it clear... this isn't a HUGE performance hit it's |
If you run profileSql you can see it... 3ms hit per attempt.
Kevin
--
Please reply using PGP.
http://peerfear.org/pubkey.asc
NewsMonster - http://www.newsmonster.org/
Kevin A. Burton, Location - San Francisco, CA, Cell - 415.595.9965
AIM/YIM - sfburtonator, Web - http://peerfear.org/
GPG fingerprint: 5FB2 F3E2 760E 70A8 6174 D393 E84D 8D04 99F1 4412
IRC - freenode.net #infoanarchy | #p2p-hackers | #newsmonster
--
MySQL Java Mailing List
For list archives: http://lists.mysql.com/java
To unsubscribe: http://lists.mysql.com/java?unsub=my...ie.nctu.edu.tw
|
« Previous Thread
|
Next Thread »
| Thread Tools | |
| Display Modes | |
Linear Mode
| |
Powered by vBulletin Version 3.5.3
Copyright ©2000 - 2013, Jelsoft Enterprises Ltd.
Copyright ©2000 - 2013, Jelsoft Enterprises Ltd.