db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Samuel Andrew McIntyre (JIRA)" <derby-...@db.apache.org>
Subject [jira] Closed: (DERBY-108) Document performance tip for inserts in tuning guide
Date Thu, 02 Jun 2005 17:22:58 GMT
     [ http://issues.apache.org/jira/browse/DERBY-108?page=all ]
     
Samuel Andrew McIntyre closed DERBY-108:
----------------------------------------

      Assign To: Jeff Levitt
     Resolution: Fixed
    Fix Version: 10.1.0.0
                     (was: 10.0.2.2)

Confirmed that the new tuning suggestion is in the Tuning Guide for 10.1.

> Document performance tip for inserts in tuning guide
> ----------------------------------------------------
>
>          Key: DERBY-108
>          URL: http://issues.apache.org/jira/browse/DERBY-108
>      Project: Derby
>         Type: Improvement
>   Components: Documentation
>     Versions: 10.0.2.1
>     Reporter: Sunitha Kambhampati
>     Assignee: Jeff Levitt
>     Priority: Minor
>      Fix For: 10.1.0.0

>
> Need to add this important performance tip in the tuning guide document.
> Could probably be added as Tip #8 in http://incubator.apache.org/derby/manuals/tuning/perf20.html#HDRSII-PERF-25864
> ----------------------------------------------
> Avoid inserts in autocommit mode if possible:
> Inserts can be painfully slow in autocommit mode. The reason is that each commit involves
a flush of the log to the disk  for each insert statement. The commit will not return until
a physical disk write has been executed. 
> To speed things up, there are 2 possibilities
> - Run in autocommit false mode and execute a number of inserts in one transaction and
then explicitly issue a commit. 
> - If your application allows an initial load into the table, one can use the import system
procedures to insert data into a table. 
> http://incubator.apache.org/derby/manuals/tools/tools90.html#HDRSII-IMPORT-57005.
> If loading into an empty table using these interfaces, derby will not log the
> individual inserts. The reason is  recovery understands that a backout of this
> operation is an empty table, and the data in the table is forced to
> disk before the transaction commits.
> --------------------------------------------
> References from derby mailing lists:
> http://nagoya.apache.org/eyebrowse/ReadMsg?listName=derby-user@db.apache.org&msgNo=363
> http://nagoya.apache.org/eyebrowse/ReadMsg?listId=271&msgNo=335

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
   http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira


Mime
View raw message