Thanks for your kind reply. I tried using both sqlentityprocessor and set
batchSize to -1but didn't get any improvement. It'd be helpful if I can see
data import handler's log.

On Saturday, November 7, 2015, Alexandre Rafalovitch <arafa...@gmail.com>
wrote:

> LoL. Of course I meant SolrJ. I had to misspell the most important
> word of the hundreds I wrote in this thread :-)
>
> Thank you Erick for the correction.
> ----
> Solr Analyzers, Tokenizers, Filters, URPs and even a newsletter:
> http://www.solr-start.com/
>
>
> On 7 November 2015 at 19:18, Erick Erickson <erickerick...@gmail.com
> <javascript:;>> wrote:
> > Alexandre, did you mean SolrJ?
> >
> > Here's a way to get started
> > https://lucidworks.com/blog/2012/02/14/indexing-with-solrj/
> >
> > Best,
> > Erick
> >
> > On Sat, Nov 7, 2015 at 2:22 PM, Alexandre Rafalovitch
> > <arafa...@gmail.com <javascript:;>> wrote:
> >> Have you thought of just using Solr. Might be faster than
> troubleshooting
> >> DIH for complex scenarios.
> >> On 7 Nov 2015 3:39 pm, "Yangrui Guo" <guoyang...@gmail.com
> <javascript:;>> wrote:
> >>
> >>> I found multiple strange things besides the slowness. I performed
> count(*)
> >>> in MySQL but only one-fifth of the records were imported. Also
> sometimes
> >>> dataimporthandler  either doesn't import at all or only imports a
> portion
> >>> of the table. How can I debug the importer?
> >>>
> >>> On Saturday, November 7, 2015, Yangrui Guo <guoyang...@gmail.com
> <javascript:;>> wrote:
> >>>
> >>> > I just realized that not everything was ok. Three child entities
> were not
> >>> > imported. Had set batchSize to -1 but again solr was stuck :(
> >>> >
> >>> > On Fri, Nov 6, 2015 at 3:11 PM, Yangrui Guo <guoyang...@gmail.com
> <javascript:;>
> >>> > <javascript:_e(%7B%7D,'cvml','guoyang...@gmail.com <javascript:;>');>>
> wrote:
> >>> >
> >>> >> Thanks for the reply. I just removed CacheKeyLookUp and CachedKey
> and
> >>> >> used WHERE clause instead. Everything works fine now.
> >>> >>
> >>> >> Yangrui
> >>> >>
> >>> >>
> >>> >> On Friday, November 6, 2015, Shawn Heisey <apa...@elyograg.org
> <javascript:;>
> >>> >> <javascript:_e(%7B%7D,'cvml','apa...@elyograg.org <javascript:;>');>>
> wrote:
> >>> >>
> >>> >>> On 11/6/2015 10:32 AM, Yangrui Guo wrote:
> >>> >>> > <entity name="movie_actress" transformer="RegexTransformer"
> >>> >>>
> >>> >>> There's a good chance that JDBC is trying to read the entire
> result set
> >>> >>> (all three million rows) into memory before sending any of that
> info to
> >>> >>> Solr.
> >>> >>>
> >>> >>> Set the batchSize to -1 for MySQL so that it will stream results to
> >>> Solr
> >>> >>> as soon as they are available, and not wait for all of them.
> Here's
> >>> >>> more info on the situation, which frequently causes OutOfMemory
> >>> problems
> >>> >>> for users:
> >>> >>>
> >>> >>>
> >>> >>>
> >>>
> http://wiki.apache.org/solr/DataImportHandlerFaq?highlight=%28mysql%29|%28batchsize%29#I.27m_using_DataImportHandler_with_a_MySQL_database._My_table_is_huge_and_DataImportHandler_is_going_out_of_memory._Why_does_DataImportHandler_bring_everything_to_memory.3F
> >>> >>> <
> >>>
> http://wiki.apache.org/solr/DataImportHandlerFaq?highlight=%28mysql%29%7C%28batchsize%29#I.27m_using_DataImportHandler_with_a_MySQL_database._My_table_is_huge_and_DataImportHandler_is_going_out_of_memory._Why_does_DataImportHandler_bring_everything_to_memory.3F
> >>> >
> >>> >>>
> >>> >>>
> >>> >>> Thanks,
> >>> >>> Shawn
> >>> >>>
> >>> >>>
> >>> >
> >>>
>

Reply via email to