On 4/5/06, Michael Levy <[EMAIL PROTECTED]> wrote:
> I am still having trouble getting Solr's war file to run under Solaris.
> I just took the most recent nightly build war, put it under /webapps/,
> point browser to /solr/admin/, and see the output below. This is a new
> install of Solaris but al
Because Solr systems are designed to be updated frequently this would make the
deleteByQuery more interesting. I guess it too could be run in parallel
assuming the index was partitioned in a round robin fashion. Any ideas? Would
this destroy performance or not be a big deal? If we assume a 5
Thanks for the hint about the score, that works well.
- Original Message
From: Mark Backman <[EMAIL PROTECTED]>
To: solr-user@lucene.apache.org
Sent: Wednesday, April 5, 2006 3:44:07 PM
Subject: Re: Solr Multisearcher
>> Does the current Solr protocol 2.0 return a score?
Yes.
Solr re
On 4/5/06, jason rutherglen <[EMAIL PROTECTED]> wrote:
> I am still interested in doing a bit more to make the replication faster.
> Rsync scans over all the files for changes which on a large index can take
> time.
We set up rsync to only look at the file size and timestamp (not
checksum it..
This looks like the bit of code to borrow from in MultiSearcher. I will mess
with implementing as a parallel multithreaded system using the
java.util.concurrent code. Seems like Solr could be run as is, meaning leaving
the caching on the index machines, have a master and at least two slaves.
> scores may be skewed, if one half of your data set has
higher occurances of certain words or phrases than the
other half.
I think that would be fine. The score would be used to narrow the results down
to the max matches.
- Original Message
From: Mark Backman <[EMAIL PROTECTED]>
To:
>> Does the current Solr protocol 2.0 return a score?
Yes.
Solr returns a score, but only if you specify it in
the field list of a query (i.e.: &fl=*,score)... there
is a bit of a performance hit for returning it, so it
is not on by default.
However, remember that this score is composed of ma
I think it would be best to write something from the ground up using the idea
of multisearcher, which would use the XML HTTP client code to obtain results
from each server. Seems like it would just do a merge. Does the current Solr
protocol 2.0 return a score? I was thinking of adding this.
I was curious as to how the caching would be handled. In a world where there
is a massive realtime virtual index, should the sorting and caching be handled
on dedicated servers separate from the index machines?
- Original Message
From: Chris Hostetter <[EMAIL PROTECTED]>
To: solr-use
I am still having trouble getting Solr's war file to run under Solaris.
I just took the most recent nightly build war, put it under /webapps/,
point browser to /solr/admin/, and see the output below. This is a new
install of Solaris but all the Tomcat test pages work and Nutch seems to
have i
: > but the first step would probably be to provide the same
: > level of functionality MultiSearcher
:
: Ahh, I was thinking the first step would be to try and use
: MultiSearcher via RemoteSearcher/RemoteSearchable.
Ah ... you were thinking something like this...
MultiSearcher(Searchable[])
On 4/5/06, Chris Hostetter <[EMAIL PROTECTED]> wrote:
> but the first step would probably be to provide the same
> level of functionality MultiSearcher
Ahh, I was thinking the first step would be to try and use
MultiSearcher via RemoteSearcher/RemoteSearchable.
-Yonik
: Lucene's MultiSearcher can already get you the top "n" documents,
: filtered, and sorted by score or other criteria.
right, that logic could probably be refactored into a base class so both
MultiSearcher(Searchable[]) and some new SolrMultiSearcher(URL[]) could
use them ... my question is more
On 4/5/06, Chris Hostetter <[EMAIL PROTECTED]> wrote:
> : IMO, it's probably a bad idea unless your index is simply too large to
> : service queries in a reasonable amount of time. Things become much
> : harder when distributed (for example, how would you take the
> : intersection of two sets of d
: IMO, it's probably a bad idea unless your index is simply too large to
: service queries in a reasonable amount of time. Things become much
: harder when distributed (for example, how would you take the
: intersection of two sets of documents in a distributed manner)? Any
: type of distributed
On 4/5/06, jason rutherglen <[EMAIL PROTECTED]> wrote:
> Is anyone looking at partitioning Solr indices over several machines and then
> querying using a multisearcher? Or creating a multisearcher like class for
> Solr? Would this idea be considered with the design of Solr or a bad idea?
>
It'
Is anyone looking at partitioning Solr indices over several machines and then
querying using a multisearcher? Or creating a multisearcher like class for
Solr? Would this idea be considered with the design of Solr or a bad idea?
17 matches
Mail list logo
