seems like there's a memleak during full reindexing

GoogleCodeExporter commented 9 years ago

I have a table with ~500.000 records. When I'm going to perform a full index 
rebuild the python process get up to 2GB of RAM at the end of the process.

The Djapian index for the model contains only 2 fields and 1 tag (id of the 
referenced model):

{{{
class ProductIndexer(WeightenedIndexer):
    fields = ('name', 'comments',
              )
    tags = (('client_id', 'client.id'),
            )
    trigger = lambda indexer, obj: obj.client and obj.client.is_enabled

    weight = 'client.registration'
}}}

The version of Djapian used: 2.3.1

Original issue reported on code.google.com by esizi...@gmail.com on 14 Feb 2010 at 12:52

GoogleCodeExporter commented 9 years ago

Do you have DEBUG=True in your settings?

Original comment by daevaorn on 14 Feb 2010 at 6:55

GoogleCodeExporter commented 9 years ago

Yes, I do. Is that a known issue?

Original comment by esizi...@gmail.com on 14 Feb 2010 at 7:01

GoogleCodeExporter commented 9 years ago

Is isn't issue -- it is Django's feature!:-) When debug is enabled Django 
stores each
db query in special log. This log can grow a lot during indexing process.

Original comment by daevaorn on 14 Feb 2010 at 7:11

Changed state: Invalid

pombreda / djapian

seems like there's a memleak during full reindexing #101