Erick,
Autocommit is commented out in solrconfig.xml. I have avoided them
until after the indexing process is complete. As an experiment I tried
committing every n records processed to see if varying n would make a
difference, it really didn't change much.
My original use case had the client running from the Solr server and
streaming the document content over from a web server based on the URL
gathered by a query from a backend database. The locking problem
appeared there first so I tried moving the client code to the web server
to be closer the the documents origin. That helped a little but ended
up locking which is where I am now.
Solr should be able to index way more documents than the 35K I'm trying
to index. It seems from other's accounts they are able to do what I'm
trying to do successfully. Therefore I believe I must be doing
something extraordinarily dumb. I'll be happy to share any information
about my environment or configuration if it will help find my error.
Thanks for all of your help.
- Tod
On 11/15/2011 8:08 PM, Erick Erickson wrote:
That's odd. What are your autocommit parameters? And are you either
committing or optimizing as part of your program? I'd bump the
autocommit parameters up and NOT commit (or optimize) from your
client if you are....
Best
Erick
On Tue, Nov 15, 2011 at 2:17 PM, Tod<listac...@gmail.com> wrote:
Otis,
The files are only part of the payload. The supporting metadata exists in a
database. I'm pulling that information, as well as the name and location of
the file, from the database and then sending it to a remote Solr instance to
be indexed.
I've heard Solr would prefer to get documents it needs to index in chunks
rather than one at a time as I'm doing now. The one at a time approach is
locking up the Solr server at around 700 entries. My thought was if I could
chunk them in a batch at a time the lockup will stop and indexing
performance would improve.
Thanks - Tod
On 11/15/2011 12:13 PM, Otis Gospodnetic wrote:
Hi,
How about just concatenating your files into one? �Would that work for
you?
Otis
----
Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
Lucene ecosystem search :: http://search-lucene.com/
________________________________
From: Tod<listac...@gmail.com>
To: solr-user@lucene.apache.org
Sent: Monday, November 14, 2011 4:24 PM
Subject: Help! - ContentStreamUpdateRequest
Could someone take a look at this page:
http://wiki.apache.org/solr/ContentStreamUpdateRequestExample
... and tell me what code changes I would need to make to be able to
stream a LOT of files at once rather than just one?� It has to be something
simple like a collection of some sort but I just can't get it figured out.�
Maybe I'm using the wrong class altogether?
TIA