You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by al...@aim.com on 2013/08/05 08:02:03 UTC

Re: additional requests sent to solr

Hello,

I still have this issue. Basically in distributed mode, when facet is true, solr-4.2 issues an additional query with 
facet.field={!terms%3D$company__terms}company&isShard=true} where for example 
company__terms have all values from company facet field.

I have added terms=false to the original query sent to solr, but it did not help.

Does anyone has any idea how to suppress these queries. 

Thanks.
Alex.


 

 

 

-----Original Message-----
From: alxsss <al...@aim.com>
To: solr-user <so...@lucene.apache.org>
Sent: Fri, Jul 19, 2013 5:00 am
Subject: additional requests sent to solr


Hello,

I send to solr( to server1 in the cluster of two servers) the folowing request

http://server1:8983/solr/mycollection/select?q=alex&wt=xml&defType=edismax&facet.field=school&facet.field=company&facet=true&facet.limit=10&facet.mincount=1&qf=school_txt+company_txt+name&shards=server1:8983/solr/mycollection,server2.com:8983/solr/mycollection

I see in the logs 2 additional requests

INFO: [mycollection] webapp=/solr path=/select params={facet=true&f.company.facet.limit=25&qf=school_txt+company_txt+name&distrib=false&wt=javabin&version=2&rows=10&defType=edismax&f.school_facet.facet.limit=25&NOW=1374191542130&shard.url=server1:8983/solr/mycollection&fl=id,score&start=0&q=alex&facet.field=school&facet.field=company&isShard=true&fsv=true} 
hits=9118 status=0 QTime=72

Jul 18, 2013 4:52:22 PM org.apache.solr.core.SolrCore execute
INFO: [mycollection] webapp=/solr path=/select params={facet=true&facet.mincount=1&company__terms=Google&ids=957642543183429632,957841245982425088,67612781366,56659036467,50875569066,957707339232706560,465078975511&facet.limit=10&qf=school_txt+company_txt+name&distrib=false&wt=javabin&version=2&rows=10&defType=edismax&NOW=1374191542130&shard.url=server1:8983/solr/mycollection&school__terms=Michigan+State+University,Brigham+Young+University,Northeastern+University&q=alex&facet.field={!terms%3D$school__terms}school&facet.field={!terms%3D$company__terms}company&isShard=true} 
status=0 QTime=6

Jul 18, 2013 4:52:22 PM org.apache.solr.core.SolrCore execute
INFO: [mycollection] webapp=/solr path=/select params={facet=true&shards=server1.prod.mylife.com:8983/solr/mycollection,server2:8983/solr/mycollection&facet.mincount=1&q=alex&facet.limit=10&qf=school_txt+company_txt+name&facet.field=school&facet.field=company&wt=xml&defType=edismax} 
hits=97262 status=0 QTime=168


I can understand that the first and the third log records are related to the 
above request, but cannot inderstand where the second log comes from. 
I see in it, company__terms and 
{!terms%3D$school__terms}school&facet.field={!terms%3D$company__terms}, whish 
seems does not have anything to do with the initial request. This is solr-4.2.0


Any ideas about it are welcome.

Thanks in advance.
Alex.

 

Re: additional requests sent to solr

Posted by alxsss <al...@aim.com>.
Hi,

Could someone please confirm that this must me so or this is a bug in SOLR.

In short, I see three logs in SOLR for one  request
http://server1:8983/solr/mycollection/select?q=alex&wt=xml&defType=edismax&facet.field=school&facet.field=company&facet=true&facet.limit=10&facet.mincount=1&qf=school_txt+company_txt+name&shards=server1:8983/solr/mycollection,server2.com:8983/solr/mycollection

for the case when facet=true.  

The third log looks like as 
INFO: [mycollection] webapp=/solr path=/select

params={facet=true&facet.mincount=1&company__terms=Google&ids=957642543183429632,957841245982425088,67612781366,56659036467,50875569066,957707339232706560,465078975511&facet.limit=10&qf=school_txt+company_txt+name&distrib=false&wt=javabin&version=2&rows=10&defType=edismax&NOW=1374191542130&shard.url=server1:8983/solr/mycollection&school__terms=Michigan+State+University,Brigham+Young+University,Northeastern+University&q=alex&facet.field={!terms%3D$school__terms}school&facet.field={!terms%3D$company__terms}company&isShard=true}
 status=0 QTime=6

where company__terms and school_terms values are taken from facet values for
company and school fields.

When data is big this leads to a log with all facet values, that
considerably slows performance. This issue is observed in distributed mode
only.

Thanks in advance.
Alex.




--
View this message in context: http://lucene.472066.n3.nabble.com/additional-requests-sent-to-solr-tp4079007p4083799.html
Sent from the Solr - User mailing list archive at Nabble.com.

Re: additional requests sent to solr

Posted by al...@aim.com.
I care about performance. Since the data is too big the query with terms becomes to long and slows performance. 

bq
---
In general distributed searchrequires two round trips to the "other" shards."
---

In this case I have three queries to solr. The third one is with {!terms..., which I do not understand why it is there.

Thanks.
Alex.

 

 

 

-----Original Message-----
From: Erick Erickson <er...@gmail.com>
To: solr-user <so...@lucene.apache.org>
Sent: Mon, Aug 5, 2013 7:10 pm
Subject: Re: additional requests sent to solr


Why do you care? Is this causing you trouble? In general distributed search
requires two round trips to the "other" shards. The first query gets the
top N, those are returned to the originator (just a list of IDs and sort
criteria,
often score). The originator then assembles the final top N, but then
the actual body of those documents must be fetched from the other
nodes.

Best
Erick


On Mon, Aug 5, 2013 at 2:02 AM, <al...@aim.com> wrote:

> Hello,
>
> I still have this issue. Basically in distributed mode, when facet is
> true, solr-4.2 issues an additional query with
> facet.field={!terms%3D$company__terms}company&isShard=true} where for
> example
> company__terms have all values from company facet field.
>
> I have added terms=false to the original query sent to solr, but it did
> not help.
>
> Does anyone has any idea how to suppress these queries.
>
> Thanks.
> Alex.
>
>
>
>
>
>
>
>
> -----Original Message-----
> From: alxsss <al...@aim.com>
> To: solr-user <so...@lucene.apache.org>
> Sent: Fri, Jul 19, 2013 5:00 am
> Subject: additional requests sent to solr
>
>
> Hello,
>
> I send to solr( to server1 in the cluster of two servers) the folowing
> request
>
>
> http://server1:8983/solr/mycollection/select?q=alex&wt=xml&defType=edismax&facet.field=school&facet.field=company&facet=true&facet.limit=10&facet.mincount=1&qf=school_txt+company_txt+name&shards=server1:8983/solr/mycollection,server2.com:8983/solr/mycollection
>
> I see in the logs 2 additional requests
>
> INFO: [mycollection] webapp=/solr path=/select
> params={facet=true&f.company.facet.limit=25&qf=school_txt+company_txt+name&distrib=false&wt=javabin&version=2&rows=10&defType=edismax&f.school_facet.facet.limit=25&NOW=1374191542130&shard.url=server1:8983/solr/mycollection&fl=id,score&start=0&q=alex&facet.field=school&facet.field=company&isShard=true&fsv=true}
> hits=9118 status=0 QTime=72
>
> Jul 18, 2013 4:52:22 PM org.apache.solr.core.SolrCore execute
> INFO: [mycollection] webapp=/solr path=/select
> params={facet=true&facet.mincount=1&company__terms=Google&ids=957642543183429632,957841245982425088,67612781366,56659036467,50875569066,957707339232706560,465078975511&facet.limit=10&qf=school_txt+company_txt+name&distrib=false&wt=javabin&version=2&rows=10&defType=edismax&NOW=1374191542130&shard.url=server1:8983/solr/mycollection&school__terms=Michigan+State+University,Brigham+Young+University,Northeastern+University&q=alex&facet.field={!terms%3D$school__terms}school&facet.field={!terms%3D$company__terms}company&isShard=true}
> status=0 QTime=6
>
> Jul 18, 2013 4:52:22 PM org.apache.solr.core.SolrCore execute
> INFO: [mycollection] webapp=/solr path=/select params={facet=true&shards=
> server1.prod.mylife.com:8983/solr/mycollection,server2:8983/solr/mycollection&facet.mincount=1&q=alex&facet.limit=10&qf=school_txt+company_txt+name&facet.field=school&facet.field=company&wt=xml&defType=edismax
> }
> hits=97262 status=0 QTime=168
>
>
> I can understand that the first and the third log records are related to
> the
> above request, but cannot inderstand where the second log comes from.
> I see in it, company__terms and
> {!terms%3D$school__terms}school&facet.field={!terms%3D$company__terms},
> whish
> seems does not have anything to do with the initial request. This is
> solr-4.2.0
>
>
> Any ideas about it are welcome.
>
> Thanks in advance.
> Alex.
>
>
>

 

Re: additional requests sent to solr

Posted by Erick Erickson <er...@gmail.com>.
Why do you care? Is this causing you trouble? In general distributed search
requires two round trips to the "other" shards. The first query gets the
top N, those are returned to the originator (just a list of IDs and sort
criteria,
often score). The originator then assembles the final top N, but then
the actual body of those documents must be fetched from the other
nodes.

Best
Erick


On Mon, Aug 5, 2013 at 2:02 AM, <al...@aim.com> wrote:

> Hello,
>
> I still have this issue. Basically in distributed mode, when facet is
> true, solr-4.2 issues an additional query with
> facet.field={!terms%3D$company__terms}company&isShard=true} where for
> example
> company__terms have all values from company facet field.
>
> I have added terms=false to the original query sent to solr, but it did
> not help.
>
> Does anyone has any idea how to suppress these queries.
>
> Thanks.
> Alex.
>
>
>
>
>
>
>
>
> -----Original Message-----
> From: alxsss <al...@aim.com>
> To: solr-user <so...@lucene.apache.org>
> Sent: Fri, Jul 19, 2013 5:00 am
> Subject: additional requests sent to solr
>
>
> Hello,
>
> I send to solr( to server1 in the cluster of two servers) the folowing
> request
>
>
> http://server1:8983/solr/mycollection/select?q=alex&wt=xml&defType=edismax&facet.field=school&facet.field=company&facet=true&facet.limit=10&facet.mincount=1&qf=school_txt+company_txt+name&shards=server1:8983/solr/mycollection,server2.com:8983/solr/mycollection
>
> I see in the logs 2 additional requests
>
> INFO: [mycollection] webapp=/solr path=/select
> params={facet=true&f.company.facet.limit=25&qf=school_txt+company_txt+name&distrib=false&wt=javabin&version=2&rows=10&defType=edismax&f.school_facet.facet.limit=25&NOW=1374191542130&shard.url=server1:8983/solr/mycollection&fl=id,score&start=0&q=alex&facet.field=school&facet.field=company&isShard=true&fsv=true}
> hits=9118 status=0 QTime=72
>
> Jul 18, 2013 4:52:22 PM org.apache.solr.core.SolrCore execute
> INFO: [mycollection] webapp=/solr path=/select
> params={facet=true&facet.mincount=1&company__terms=Google&ids=957642543183429632,957841245982425088,67612781366,56659036467,50875569066,957707339232706560,465078975511&facet.limit=10&qf=school_txt+company_txt+name&distrib=false&wt=javabin&version=2&rows=10&defType=edismax&NOW=1374191542130&shard.url=server1:8983/solr/mycollection&school__terms=Michigan+State+University,Brigham+Young+University,Northeastern+University&q=alex&facet.field={!terms%3D$school__terms}school&facet.field={!terms%3D$company__terms}company&isShard=true}
> status=0 QTime=6
>
> Jul 18, 2013 4:52:22 PM org.apache.solr.core.SolrCore execute
> INFO: [mycollection] webapp=/solr path=/select params={facet=true&shards=
> server1.prod.mylife.com:8983/solr/mycollection,server2:8983/solr/mycollection&facet.mincount=1&q=alex&facet.limit=10&qf=school_txt+company_txt+name&facet.field=school&facet.field=company&wt=xml&defType=edismax
> }
> hits=97262 status=0 QTime=168
>
>
> I can understand that the first and the third log records are related to
> the
> above request, but cannot inderstand where the second log comes from.
> I see in it, company__terms and
> {!terms%3D$school__terms}school&facet.field={!terms%3D$company__terms},
> whish
> seems does not have anything to do with the initial request. This is
> solr-4.2.0
>
>
> Any ideas about it are welcome.
>
> Thanks in advance.
> Alex.
>
>
>