Re: Optimize cascade deletes

From: Giulio Cesare Solaroli (giulio.cesar..mail.com)
Date: Sun Sep 23 2007 - 03:38:00 EDT

  • Next message: Aristedes Maniatis: "Re: Transactions , Primary Keys"

    Hello Aristedes,

    On 9/23/07, Aristedes Maniatis <ar..aniatis.org> wrote:
    >
    > On 22/09/2007, at 11:41 PM, Giulio Cesare Solaroli wrote:
    >
    > > 15 seconds are a lot of time, but they would be bearable. The problem
    > > is that I suppose this time will gradually increase with the growth of
    > > the number of records to delete.
    >
    > That is still about 1000 times the duration I'd expect this to take.
    > There is still a significant problem which you might be able to
    > identify with some profiling or logging.

    [15:21:12] DELETE FROM clipperz.RCRVRS WHERE ID_RCRVRS = ?
    [15:21:12] [batch bind: 1433]
    [15:21:12] [batch bind: 797]
    [....]
    [15:21:12] [batch bind: 55637]
    [15:21:12] [batch bind: 51363]
    [15:21:27] === updated 182 rows.

    the above log was generated using these log4j preferencess:

    log4j.logger.org.apache.cayenne=ERROR
    log4j.logger.org.apache.cayenne.access.QueryLogger=DEBUG

    Is there anything going on on the Cayenne code between the last [batch
    bind:xxx] log and the "updated" log?

    Is all this time spent on the DB only?

    Thanks for your feedback.

    Best regards,

    Giulio Cesare

    This archive was generated by hypermail 2.0.0 : Sun Sep 23 2007 - 03:38:34 EDT