You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-user@lucene.apache.org by Scott Smith <SS...@MainstreamData.com> on 2003/12/12 22:25:16 UTC
Lucene Art
I'm not having a problem. The question is whether I picked a reasonable set
of parameters for what I'm doing.
I have an application which receives messages. Each message averages around
4k bytes and I get, on average, 0-10 every minute. So my application simply
wakes up once each minute, opens an IndexWriter, creates Docs for each
message, adds them to the index, and then closes the index writer. Once per
day I optimize the index. The access to the index is via webpages (so
multiple simultaneous accesses can occur and I don't want too much latency
between a message arriving and it's availability in the index). Based on
posts I saw in the user archives, I set the mergeFactor to 10 (which I think
is the default anyway). I probably get 10-20 MB's per day and I'll want to
index 60 day's of messages (so 600MB to 1.2GB of messages).
Does this sound like a reasonable approach? Any parameters I should change?
Any others I should be setting? So far it is working fine, but I haven't
had enough messages to see how things perform when the index gets large and
I am new to using Lucene. Wondered what people with more experience with
Lucene thought. I'm using 1.3RC3 if it matters.
Scott
---------------------------------------------------------------------
To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: lucene-user-help@jakarta.apache.org
Re: Lucene Art
Posted by Doug Cutting <cu...@lucene.com>.
Your setup sounds good to me.
Scott Smith wrote:
> I'm not having a problem. The question is whether I picked a reasonable set
> of parameters for what I'm doing.
>
> I have an application which receives messages. Each message averages around
> 4k bytes and I get, on average, 0-10 every minute. So my application simply
> wakes up once each minute, opens an IndexWriter, creates Docs for each
> message, adds them to the index, and then closes the index writer. Once per
> day I optimize the index. The access to the index is via webpages (so
> multiple simultaneous accesses can occur and I don't want too much latency
> between a message arriving and it's availability in the index). Based on
> posts I saw in the user archives, I set the mergeFactor to 10 (which I think
> is the default anyway). I probably get 10-20 MB's per day and I'll want to
> index 60 day's of messages (so 600MB to 1.2GB of messages).
>
> Does this sound like a reasonable approach? Any parameters I should change?
> Any others I should be setting? So far it is working fine, but I haven't
> had enough messages to see how things perform when the index gets large and
> I am new to using Lucene. Wondered what people with more experience with
> Lucene thought. I'm using 1.3RC3 if it matters.
>
> Scott
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: lucene-user-help@jakarta.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: lucene-user-help@jakarta.apache.org