LucidWorks support Forum » Recent Posts

Sami Siren on "Mutually Authenticated SSL (NeedClientAuth)"

Sami Siren — Sun, 05 Aug 2012 23:38:55 +0000

Hi Dan,

Yes it is possible to run core with mutually authenticated ssl. Since core also communicates to core with https there are couple of additional knobs you need to tweak so that the internal clients send the appropriate client certificate. Please see http://lucidworks.lucidimagination.com/display/lweug/SSL+Configuration for more information.

Lance on "java processes do not start after computer restart"

Lance — Mon, 30 Jul 2012 19:41:06 +0000

Is this a Unix or Windows computer? If it is Unix, you need to create a service script that starts LWE on startup.

If it is Windows, the installer sets the programs to restart when you log in.

Lance on "java processes do not start after computer restart"

Lance — Mon, 30 Jul 2012 19:40:41 +0000

Is this a Unix or Windows computer? If it is Unix, you need to create a service script that starts LWE on startup.

If it is Windows, the installer sets the programs to restart when you log in.

Erik Hatcher on "LWE slowness,"

Erik Hatcher — Sun, 29 Jul 2012 06:16:51 +0000

This is a duplicate post.

It's probably the business rules integration, which has a known indexing performance hit. If you're not using that feature, you can disable it: http://lucidworks.lucidimagination.com/display/lweug/Disabling+Business+Rules

Erik Hatcher on "LWE slowness,"

Erik Hatcher — Sun, 29 Jul 2012 06:16:06 +0000

mersad.p on "LWE slowness,"

mersad.p — Sat, 28 Jul 2012 21:56:20 +0000

I am wondering why LWE is so solwer that solr, even the indexing time is much more, though it displays a shorter time failsely.

mersad.p on "LWE slowness,"

mersad.p — Sat, 28 Jul 2012 21:54:49 +0000

I am wondering why LWE is so solwer that solr, even the indexing time is much more, though it displays a shorter time failsely.

mersad.p on "java processes do not start after computer restart"

mersad.p — Sat, 28 Jul 2012 21:52:34 +0000

When the computer is restated, the two main java proceses do not start and the server dose not work so that the lwe pages are not availble. Can anay body help?

denis on "UI error"

denis — Fri, 27 Jul 2012 11:25:43 +0000

Thanks, Pascal. This issue was fixed in LucidWorks Enterprise 2.1.

pascal.essiembre on "UI error"

pascal.essiembre — Fri, 27 Jul 2012 11:04:52 +0000

In case others have the problem, we found out the fields page would not render if any of the fields (dynamic ones in our case) had a dot in their name (like "metatag.description"). Probably happens with other characters too. Not sure yet if this has been fixed in releases more recent than 2.0, but I am sharing this info in case it helps others.

b.lanoy.10 on "LucidWorks Enterprise JDBC problems"

b.lanoy.10 — Fri, 27 Jul 2012 07:47:30 +0000

Hello,

I've problem with LWE when I try to create new user just for testing what the lambda user can see I've the internal server error

the log say that :

Started POST "/users" for 127.0.0.1 at Fri Jul 27 13:05:42 UTC 2012
2012-07-27 13:05:42,111 INFO jruby.lwe-ui - Processing by UsersController#create as HTML
2012-07-27 13:05:42,111 INFO jruby.lwe-ui - Parameters: {"utf8"=>"✓", "authenticity_token"=>"f7dnUYuzyb4urrDybG9l9epj9H5ImHXvqe5g2SI08uA=", "user"=>{"username"=>"test00", "password"=>"[FILTERED]", "password_confirmation"=>"[FILTERED]", "email"=>"", "authorization"=>"search", "timezone"=>"UTC"}, "commit"=>"Create User"}
2012-07-27 13:05:42,283 INFO jruby.lwe-ui - Completed 500 Internal Server Error in 172ms
2012-07-27 13:05:42,330 FATAL jruby.lwe-ui -
ActiveRecord::JDBCError (attempt to write a readonly database):
app/controllers/users_controller.rb:18:in create'   app/controllers/core/base_controller.rb:35:inlong_solr_requests_timeout'
app/controllers/core/base_controller.rb:33:in `long_solr_requests_timeout'

believe that is because I try to write on read only database but I'm not sure and I don't know why the log say that. Do you know why it's happened ? I'm quit lost...
Thx in advance

mersad.p on "LucidWorks Enterprise JDBC problems"

mersad.p — Tue, 24 Jul 2012 03:03:45 +0000

Yes, I mean the built-in sarch page in LWE . I tried different types like "string" or "TextTight" or "text-en" but it dose not work with arabich while it works with persian.

Lance on "LucidWorks Enterprise JDBC problems"

Lance — Mon, 23 Jul 2012 14:52:00 +0000

When you talk about searching, do you mean with the built-in search page in LWE? Or with your own application?

Since you use a database, what fields do you use and what are their field types?

mersad.p on "LucidWorks Enterprise JDBC problems"

mersad.p — Sun, 22 Jul 2012 03:33:44 +0000

thanks so much.

yes it brings the indexed data with empty seaarch. even it bring the results when the database is in english but I get no result when the databse is in a arabic language.

what should be the problem
?

Erik Hatcher on "Handling Phrase Query"

Erik Hatcher — Fri, 20 Jul 2012 08:11:25 +0000

Looks like you'll need to tinker with the various edismax settings. Try pf2=productName&pf3=productName

Erik Hatcher on "delete some data in solr"

Erik Hatcher — Fri, 20 Jul 2012 08:05:39 +0000

Andrew - again your markup didn't come through, so I can't really make sense of your post, apologies. (our forum system is going to be replaced one of these days). But a space in an id shouldn't matter, just be sure you're specifying it exactly as it was for indexing.

andrewfiade on "delete some data in solr"

andrewfiade — Thu, 19 Jul 2012 07:35:41 +0000

ok, thanks Erick, i have try and succes my script

curl http://127.0.0.1:8888/solr/iqroit/update/?commit=true -H "Content-Type: text/xml" --data-binary ' https://apps.facebook.com/coca-colaparks/'; i run in console, some time, fail execute, cause space between with keyword

andrewfiade on "delete some data in solr"

andrewfiade — Thu, 19 Jul 2012 07:28:00 +0000

delete

Erik Hatcher on "delete some data in solr"

Erik Hatcher — Thu, 19 Jul 2012 05:33:39 +0000

Again, markup (XML and HTML tags) don't work well in our forums, so plain text is best. But looks like your delete worked fine. Be sure now to commit (you can simply add &commit=true to your /update request) and the document now should not be found.

andrewfiade on "delete some data in solr"

andrewfiade — Wed, 18 Jul 2012 22:35:07 +0000

maybe mr Erick can show me, where exaclty id from my xml. you can see and feel free to see my experiment, in my web: http://202.47.88.77:8888/solr/iqroit/select/?q=facebook&version=2.2&start=0&rows=10&indent=on i have try by id:

http://localhost:8888/solr/iqroit/update?stream.body=%3Cdelete%3E%3Cid%3Ehttps://apps.facebook.com/minimaps/%3C/id%3E%3C/delete%3E

with ouput message: --------------------------------------- This XML file does not appear to have any style information associated with it. The document tree is shown below. 0 1 --------------------------

Erik Hatcher on "delete some data in solr"

Erik Hatcher — Wed, 18 Jul 2012 08:38:58 +0000

For the example you provide, if you used the delete-by-id feature rather than by query it will work. The issue is that deleting by query requires the expression to be parsed, and because the id value itself has another ":" in it, it is an invalid expression. You can try surrounding the value by quotes such as id:"-value-", but if you're always deleting by id, use the delete-by-id rather than delete-by-query feature. See here for details and examples: http://wiki.apache.org/solr/UpdateXmlMessages#A.22delete.22_documents_by_ID_and_by_Query

andrewfiade on "delete some data in solr"

andrewfiade — Wed, 18 Jul 2012 07:58:00 +0000

this is message error:

when i run : http://202.47.88.77:8888/solr/iqroit/update?stream.body=%3Cdelete%3E%3Cquery%3Eid:https://www.facebook.com/muhammad.muzaqi%3C/query%3E%3C/delete%3E

with following error:

-----------------------------

<h2>HTTP ERROR 400</h2>

Problem accessing /solr/iqroit/update. Reason:

    Cannot parse 'id:https://www.facebook.com/muhammad.muzaqi': Encountered " ":" ": "" at line 1, column 8.
Was expecting one of:
    <EOF>
    <AND> ...
    <OR> ...
    <NOT> ...
    "+" ...
    "-" ...
    <BAREOPER> ...
    "(" ...
    "*" ...
    "^" ...
    <QUOTED> ...
    <TERM> ...
    <FUZZY_SLOP> ...
    <PREFIXTERM> ...
    <WILDTERM> ...
    <REGEXPTERM> ...
    "[" ...
    "{" ...
    <NUMBER> ...
    ------------------------------------------------------------

my analysis, because, solr use primary key id.and if we crawl website , and data uniq is address webso id recordwill be use:   http://

if we query, <delete><query>id:https://www.facebook.com/muhammad.muzaqi</query></delete>will be error, different, if id/primary key do not use http:// only use alfabhet

Erik Hatcher on "delete some data in solr"

Erik Hatcher — Wed, 18 Jul 2012 07:45:15 +0000

Our forum software must have stripped your error message. Can you please post the latest details again as plain text (no HTML or XML markup)?

andrewfiade on "delete some data in solr"

andrewfiade — Wed, 18 Jul 2012 07:20:51 +0000

thanks:

http://202.47.88.77:8888/solr/iqroit/update?stream.body=%3Cdelete%3E%3Cquery%3Eid:https://www.facebook.com/muhammad.muzaqi%3C/query%3E%3C/delete%3E

now, different message error

<div id="gt-res-content" class="almost_half_cell">
<div dir="ltr">but my prediction problem "id" to be true when I try
please try, no problem for me is removed, because it is still in the experimental

Erik Hatcher on "delete some data in solr"

Erik Hatcher — Wed, 18 Jul 2012 07:12:51 +0000

In your Solr requests you need to ensure the core name is also present. Your example that had an error did not have iqroit in the URL. The update URL should be http://202.47.88.77:8888/solr/iqroit/update...

andrewfiade on "delete some data in solr"

andrewfiade — Wed, 18 Jul 2012 06:22:16 +0000

this is my solr:

http://202.47.88.77:8888/solr/iqroit/select/?q=facebook&version=2.2&start=0&rows=10&indent=on

I deliberately show my web address, because it is still in the experimental.

my question, how we can delete some data in xml via solr.

in tutorial: http://wiki.apache.org/solr/UpdateXmlMessages;

we can delete, used :

<delete><id>05991</id></delete>
<delete><query>office:Bridgewater</query></delete>

the problem in my web, id from result xml, i think with keyword id.example in my web, xml:<str name="id">https://www.facebook.com/muhammad.muzaqi</str>where my id use symbol  :

example:http://202.47.88.77:8888/solr/update?stream.body=%3Cdelete%3E%3Cquery%3Eid:https://www.facebook.com/muhammad.muzaqi%3C/query%3E%3C/delete%3E

always, message error not found

any idea and suggestion to me

thanksiqroit.com

gourab.bhattacharya on "Handling Phrase Query"

gourab.bhattacharya — Tue, 17 Jul 2012 23:26:26 +0000

The field type of productName is text_en. Moreover, I have also tried this with edismax, and the output is same. I have also tried different Tokenizer and Filter factory in schema.xml for the field type text_en, but with no success.

Here is the output in debug mode, when edismax is used.
<lst name="debug">

<str name="parsedquery">(+((DisjunctionMaxQuery((productName:pink)) DisjunctionMaxQuery((productName:mill)) DisjunctionMaxQuery((productName:non)) DisjunctionMaxQuery(((productName:fat productName:stout productName:corpulent productName:fleshy productName:beefy productName:paunchy productName:plump productName:full productName:rotund productName:tubby productName:pudgy productName:chubby productName:chunky productName:burly productName:bulky productName:elephantine))) DisjunctionMaxQuery((productName:dry)) DisjunctionMaxQuery((productName:milk)))~6))/no_coord</str>

<str name="parsedquery_toString">+(((productName:pink) (productName:mill) (productName:non) ((productName:fat productName:stout productName:corpulent productName:fleshy productName:beefy productName:paunchy productName:plump productName:full productName:rotund productName:tubby productName:pudgy productName:chubby productName:chunky productName:burly productName:bulky productName:elephantine)) (productName:dry) (productName:milk))~6)</str>
------------------------------------------------------------------------------------------------------------------------------------

As we can see, the difference with lucid is that we dont have <PhraseQuery(productName:"non fat dry"> here, that is using edismax each word in the query string forms the phraseQuery and not the combination of words. Another difference is that using edismax, solr is also searching for the various synonyms configured for the word 'fat' in the query string.

Looking forword to your help.

Erik Hatcher on "Handling Phrase Query"

Erik Hatcher — Tue, 17 Jul 2012 10:41:46 +0000

What is the field type of productName and how is that field type configured?

I'd recommend using dismax or edismax at this point, as the "lucid" query parser has some known issues (specifically with lowercasing when it should not).

gourab.bhattacharya on "Handling Phrase Query"

gourab.bhattacharya — Tue, 17 Jul 2012 10:28:58 +0000

I am querying solr as, q=Pink Mill Non fat Dry Milk&qf=productName, its fetching one result and rightly so.

I have another data, "productName:Red Mill Non fat Dry Milk" in another document. But the document containing this is not fetched.

In debug mode, here is the output while I am using lucid as defType-

<str name="parsedquery">+productName:pink +productName:mill +productName:non +productName:fat +productName:dry +productName:milk PhraseQuery(productName:"pink mill"~50) PhraseQuery(productName:"mill non"~50) PhraseQuery(productName:"non fat"~50) PhraseQuery(productName:"fat dry"~50) PhraseQuery(productName:"dry milk"~50) PhraseQuery(productName:"pink mill non"~50) PhraseQuery(productName:"mill non fat"~50) PhraseQuery(productName:"non fat dry"~50) PhraseQuery(productName:"fat dry milk"~50)</str>

<str name="parsedquery_toString">+productName:pink +productName:mill +productName:non +productName:fat +productName:dry +productName:milk productName:"pink mill"~50 productName:"mill non"~50 productName:"non fat"~50 productName:"fat dry"~50 productName:"dry milk"~50 productName:"pink mill non"~50 productName:"mill non fat"~50 productName:"non fat dry"~50 productName:"fat dry milk"~50</str>

-----------------------------------------------------------------------------------------------------------

As we can see, one of the phrase query is <PhraseQuery(productName:"non fat dry"> which is contained in "RedMill Non fat Dry Milk". Then why that document is not returned in search result??? What configurations need to be done such that the sub string of the 'q' field is searched???

Lance on "Geospatial search"

Lance — Tue, 17 Jul 2012 08:28:23 +0000

Hi-

None of us have worked with the geospatial code. Also, it will be completely different in the next LucidWorks release.

What do you want to do with it?

ramanjaneyulut on "Geospatial search"

ramanjaneyulut — Mon, 16 Jul 2012 22:24:33 +0000

Hi,
Data is crawling (data source type website) but no data for geospatial is getting indexed. (Added new geohash filed to the data source).
While crawling the data with the given default geohash value got indexed with it.
I went through the LWE works platform documentation but could not find the geospatial field mapping. Please provide your inputs on geospatial.
 
Regards,
Ram

Lance on "LucidWorks Enterprise JDBC problems"

Lance — Mon, 16 Jul 2012 17:53:13 +0000

Do you search for a string? Try searching with an empty search box. That should load all documents.

Lwe is not necessarily slow. The UI gives this message for many kinds of errors when talking to LWE.

Erik Hatcher on "Google Connector Manager-based crawler"

Erik Hatcher — Mon, 16 Jul 2012 13:40:47 +0000

Though we will make it possible to plug in Google Connectors in a future release. (it's kind of possible now, but undocumented and requires a number of crude steps and Java glue coding)

Lance on "Google Connector Manager-based crawler"

Lance — Mon, 16 Jul 2012 13:35:49 +0000

No, we have custom code for using the Google code for Sharepoint. We do not support other connectors.

jotschgl on "Google Connector Manager-based crawler"

jotschgl — Mon, 16 Jul 2012 07:25:03 +0000

In your Doku the sharepoint crawler is called "Google Connector Manager-based crawler"

Does this mean that I can use other connectors from the the Google Search Appliance like its possible with constellio?

Heinz