lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ganesh Sethuraman <ganeshmail...@gmail.com>
Subject Re: solr cloud unique key query request is sent to all shards!
Date Mon, 19 Feb 2018 21:22:45 GMT
This works !. Both V1 and V2 version of the real time get works fine.
Just an added note, the performance (response time) also improved.

Thanks Tom

On Mon, Feb 19, 2018 at 1:17 AM, Tomas Fernandez Lobbe <tflobbe@apple.com>
wrote:

> In real-time get, the parameter name is “id”, regardless of the name of
> the unique key.
>
> The request should be in your case: http://<host>:8080/api/
> collections/col1/get?id=69749398
>
> See: https://lucene.apache.org/solr/guide/7_2/realtime-get.html
>
> Sent from my iPhone
>
> > On Feb 18, 2018, at 9:28 PM, Ganesh Sethuraman <ganeshmailbox@gmail.com>
> wrote:
> >
> > I tried this real time get on my collection using the both V1 and V2 URL
> > for real time get, but did not work!!!
> >
> > http://<host>:8080/api/collections/col1/get?myid:69749398
> >
> > it returned...
> >
> > {
> >  "doc":null}
> >
> > same issue with V1 URL as well, http://
> > <host>:8080/solr/col1/get?myid:69749398
> >
> > however if i do q=myid:69749398 with "select" request handler seems to
> > fine. I checked my schema again and it is configured correctly.  Like
> below:
> >
> > <uniqueKey>myid</uniqueKey>
> >
> > Also i see that this implicit request handler is configured correctly Any
> > thoughts, what I might be missing?
> >
> >
> >
> > On Sun, Feb 18, 2018 at 11:18 PM, Tomas Fernandez Lobbe <
> tflobbe@apple.com>
> > wrote:
> >
> >> I think real-time get should be directed to the correct shard. Try:
> >> [COLLECTION]/get?id=[YOUR_ID]
> >>
> >> Sent from my iPhone
> >>
> >>> On Feb 18, 2018, at 3:17 PM, Ganesh Sethuraman <
> ganeshmailbox@gmail.com>
> >> wrote:
> >>>
> >>> Hi
> >>>
> >>> I am using Solr 7.2.1. I have 8 shards in two nodes (two different m/c)
> >>> using Solr Cloud. The data was indexed with a unique key (default
> >> composite
> >>> id) using the CSV update handler (batch indexing). Note that I do NOT
> >> have
> >>> <prefix and an !> while indexing.   Then when I try to  query the
> >>> collection col1 based on my primary key (as below), I see that in the
> >>> 'debug' response that the query was sent to all the shards and when it
> >>> finds the document in one the shards it sends a GET FIELD to that shard
> >> to
> >>> get the data.  The problem is potentially high response time, and more
> >>> importantly scalability issue as unnecessarily all shards are being
> >> queried
> >>> to get one document (by unique key).
> >>>
> >>> http://<host>:8080/solr/col1/select?debug=true&q=id:69749278
> >>>
> >>> Is there a way to query to reach the right shard based on the has of
> the
> >>> unique key?
> >>>
> >>> Regards
> >>> Ganesh
> >>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message