In this blog we discuss a possible to the Search language translation and other issue with the Sharepoint translations

Let start with the solution to the customized refiner, for the different languages installed the farm.

For each language we should create a new site collection, type Search, Fast center template, remember choose the right language, and give a URL.

We aim our browser to the root search center; click Site Actions||Edit Page|| Add a new web part || Media and Content||Content Editor|| i.e “Top Zone”.

Edit the web part Content editor|| Deploy appearance|| give a name|| and chose Chrome type “none”. (It will make invisible the title)|| Apply||OK

Click in  add new content||in the Editing tools tab select insert||link||from Sharepoint ||chose site and the URL of the precedent Search site collection that we created, and click OK.

By clicking in Links Tools Format, we will complete the link configuration adding the total URL and description; also you can open the link in a new tab and/or display the icon, save the change and close the edit.

Publish the changes

By clicking in the new link we will redirect to the new search center in our customized language.

Now it is time to configure the refiners and the other web parts, for more info see the part 1 of this blog’s series to see how to edit refiners.

After you have deployed the refiners everything should be translated to the selected language.

A Similar case :

Iin the user web application. My site that web application when needs to be translated, some web part fail in the translation, that behavior has the same root it is a title possible to customize.

 Other workaround:

It might be you want use a webapplication instead a site collection to workaround this case, it will work just need to decleare the web applications address into the link configuration, the soluiton of Extend the web application it might be over load your systems.

I cannot blame them about this confusion, so a way to face this kind of questions is on the field.

Let’s add the Norwegian language in our farm and we won’t override the translations.

We search in English

The next steps is searching after we changed our “Select Display Language” form the user account item drop down, we select Norwegian, or other language you might wanted display.

As you can see between the two previous, the text in the squares is translated.

This is because we are using the default configuration.  Roll back to English language, and  we'll have a look to the filter category definition’s value.

Site actions || Edit page|| Edit Refinement Panel Web Part|| filter category field: copy the content and paste into any editor. We can observe the highlighted words correspond with the values in the black  squares in the English search screenshot.

Let check the Norwegian configuration and see if we see the same words in the xml than in the red squares, so change the language and perform Site actions || Edit page|| Edit Refinement Panel Web Part|| filter category field: copy the content and paste into any editor. We can observe the highlighted words correspond with the values in the red squares in the Norwegian search screenshot

Now we observed the behavior with the default configuration, let’s compare how it behaves with a customized configuration for the refiners.

To do this we need to disable the Default configuration in the Refiner web part so Site actions || Edit page|| Edit Refinement Panel Web Part|| clear the box “Use Default configuration”

Edit your filter category configuration, in my case I have insert a few modifications but I will focus in the refiner named workphone and alphabetic.

Because I define the configuration in English I don’t specs any translations get done.

Now we will submit a query in the Search center using Norwegian language, it is expected translation but

In red squares you can observe the Norwegian translations in green the Details of the customized Refiners.

The reason been for this leak of translation has the fundaments in the infinite possible combination that a single title can get, also the orthography, watch as the second refiner is misspelled Alpabetic, it should be “Alphabetic” , it will introduce a new level of complexity since Sharepoint should also check for spelling error before publish the page.

The customized refiners do not get translated, so if we want customized in a certain language, it is need customize the refiners xsl code using the wished Language words.

So let see what the result in the using Norwegian Display Language are.

Besides observe also that the tabs names and the sort by parameter are not translated, the first couple it is hardcoded when the Search site is created, the second belong to a static property, which is associated with the ranking and is case of the name changing the ranking value will be lost.

 If you are thinking to create a second default.aspx and a second result.aspx , and then handle the language issue from there I must to said that is not straight forward and there always the boundaries ofhold both language versions for the site collection, for the same user.
There is a way to work around this issue and it is create one or extend the web application and connect both web application afterwards, but this topic will be discuss in the next part of the Blog.

In this part we will see how the Search Center in Sharepoint 2010 renders the Navigator information.

By default the web part Refiners on Search center does not include the customized managed properties, only a few standard properties like “result type” “Site” or “Author”

Because we are interested in show the values of our FAST navigator that return the worknumbers, it is need to edit and modify the search center page.

By clicking in the “Site Action”, “Edit Page”, on the left Zone, click and the Refinement Panel’s arrow, Edit web part , deploy Refinement , untick  use default Configuration, select the field content, “Filter Category Defintion”, copy and paste in your favorite editor

Edit the pasted text and insert this code in the wished possition you want to show the refiner in the search center

<Category Title=”Result workphone” Description=”The phone extension of the item”

Type=”Microsoft.Office.Server.Search.WebControls.ManagedPropertyFilterGenerator” MetadataThreshold=”1” NumberOfFiltersToDisplay=”10”

MaxNumbeOfFilters=”20” ShowMoreLink=”True” MappedProperty=”skypephone” MoreLinkText=”show more” LessLinkText=”show fewer”

ShowCounts=”Count” SortBy=”NumericValue” SortDirection=”Ascending”/>

 Paste back to the Filter category definition field, click apply, OK, and save the change.

Submit the query again and the new refiners should appear, if they do not show, review your xsl code for possible misspellings.

Observe the refiners appear sorting by ascending,

NOTE:
Do not confuse the Navigator/Refiner sorting with the Search core web part "Sort By" feature, whihc sort the returned documents they are differents.

In this third part we will be focus in the sorting result, in the FAST Server by using the Qr-server.

Let take the point in the result returning when we run submit the “fasta” query using the navigator “skypephone”.

As you can see the navigator element are not sorted,

The Qrserver return the result order based in the ranking.

For example the number 123999 is in the second position because the document that contains this value is returned as hit number 2

The number 444999 is in the fourth position because the document that contains this value is returned as hit number 4

NOTE: in the two previous pictures the highlighted the hit value that belong to the next document in the result list.

So in order to make a sorting result we will manipulate the FQL GUI.

Observe the order of the numbers is ascendant, after we have introduced the symbol ‘+’ in the fql.

Let check the order of the documents when we make a sorting result descendant. We set up the next fql

NOTE: In this case the sort is alphanumeric not a sorting integers operation.

If we check the first result in the navigator value 999677889 should appear in the first position.

A further detail to note is the ranking become zero, for all the documents; let’s see with the how the ranking value change before use the sorting and after use the sorting. for example the document with phone number 123999

Without using the sorting feature

Using the sorting

NOTE:Observe the internalid value, this is the field we use to identify the documents and to know if we are watching two different documents or the same document.

In this part we will focus in the searching part, to be more specific in the querying part, using navigators .

Lets start, from a FAST server where the QR Server module is installed point your web browser to the port 13280, and submit your FQL query, perhaps "fasta" and click search.

Opps !!!! no results rendered…L an it seems an error message.

<QUERYTRANSFORM NAME=”FastQT_Lemmatizer” ACTION =“nop” QUERY=’” CUSTOM =“No change to query” MESSAGE =“Lemmatization turned off for currentquery” MESSAGEID =16” INSTANCE =“lemmatizer” I>

<QUERYTRANSFORM NAME =“ FastQT_ResubmitQuery” ACTION =“nop” QUERY =0 CUSTOM =“Resubmit off flag present, skipping resubmit.” MESSAGE =“Query not resubmitted” MESSAGEID=”2” INSTANCE=”° I>

<QUERYTRANSFORM NAME=”Final query’ ACTION=”NOP” QUERY=”and(filter(or(”docacl”:string(”all”), “docacl”:string(”unknown”)),

annotation_class”invisible”, annotation_class=”sam”), xrank(string(”fasta”), “format” :string( “unknown format”, mode “phrase”,

linguistics”off”), “format”:string(”microsoft excel”, mode”phrase”, linguistics”off”), “format”:string(”xml document”, mode”phrase”,

linguistics”off”), “fileextension”:string(”csv”, mode”phrase”, linguistics”off”), “fileextension”:string(”zip”, mode”phrase”, linguistics”off”),

“fileextension”:string(”rtf”, mode”phrase”, linguistics”off”), “fileextension”:string(”vsd”, mode”phrase”, linguistics”off”),

“fileextension”:string(”oft”, mode”phrase”, linguistics”off”), “fileextension”:string(”msg”, mode”phrase”, linquistics”oW’),

“flleextension”:string(”txt”, mode”phrase”, linguistics”off”), “isemptylist”:string(”true”, mode”phrase”, linguistics”off”), “islistitem”:string

(“true”, mode”phrase”, linguistics”off”), boostall&’no”, boostint(”-4000”)))’ CUSTOM=”FQ[.” MESSAGE=”Final query” MESSAGEID=’l° I>

Let check the every module is running.

Form A powershell windows run the command >>nctrl status , an confirm all the modules are running.

Poweshell picture

Lets override the qtf-securityfql and Disable it (just for test and debugging aftewords enable again otherwise all your documents will be public):

1. Log into each QRServer and go to $DRIVE:\FASTSearch\etc\qrserver
2. Open qtf-config.xml file.
3. Scroll to the bottom and comment out this line:<instance-ref name=”securityfql”    critical=”1″ />.
4. Restart QRServer: nctrl restart qrserver

Picture qtf disable.

Try again and submit the query, now we should get results.

If you remember in the previous part 1 of this blog serie's we were indexing a few customized properties, one of them was ows_workphone which it is turn out to be mapped to skypephone Managed property.

In order to confirm the crawling did not dismiss any document let search creating a FQL that gather the result by grouping them in one field call, skypephone . This filed will ne named Navigator

To do that, just repeat the previous query but this time adds the Navigator option by ticking the box and adding the navigator value.

and here we have the result

Observe the Navigation elements count value(13), more than all the phone number associated to the documents, in our original library (11) .

So the questions now are :

1)      Where is it the missed phone number 9994554 ?

2)      Why I have more than 12 result in the navigator

Let 's troubleshoot the search using the Qrserver.

1)      Where is gone the documents associated to 9994554?
 Submit a query with the phone number.

Jufff luckily is here, observe the title of the document is “fatsa” not “fasta” (our initial query), so the search engine dismiss this document, and the document is not included in the result

2)      If my document library has 12 documents and 1 is not included in the result, why I get 13 telephone numbers in the navigator result?

Possible answer : my system has duplicate result.

Unfortunately this is not the case, if you submit the initial query again but this time we change the navigator parameter to sitename.

And here we have the explanation.

There are 4 other document that match the fasta query in the location http://moss14/techsupport

Two of them, are documents as you can observer in the hit number.

The other is the container.

And the forth is the view (or aspx page).

If we go navigate to the url http://moss14/techsupport/shared_documents  we could confirm all these hits exist.

So it should be other explanation, observe the last screenshoot.

The explanations is,  last column with telephone number is the workphone column which after crawl should became ows_ workphone the same CP we mapped to the skypephone therefore the values for that property are included into the navigator result.

that is the reason we see 13 result, 11 documents from site http://moss14 and 2 document from http://moss14/techsupport

One way to see the problem is: the turds do not remove in the past, confuse us in the search result.

Other way is the navigators are a powerful feature that could return us unexpected result.

It is up to you how to use them.

!!!!Thanks To Nicolas U that got the patience to teach me the navigator funtinality.!!!!!

Also some tips about how to control the healthiness of the process, in order to troubleshoot the potential issues.

The first part of the blog is focus in the crawling.

The second parts focus in the search.

The third part will focus in to the sorting of rendered result.

These are the three columns details and the document library overview.

The first time the Sharepoint run a full crawl those columns will be added automatically as Crawled properties, it is possible see their details in the Property management section of the FAST Query SSA.

I will not stop to explain how to create and manage that Crawled and managed Properties, since it is documented in this link

Usually the get created as ows_ <column_name>,

You can create two different Crawled properties with the same name but different data type

Or you can map the property to more than one Managed property (see link).

So let assume you already crawl you document library and you see the generated Crawled properties.

Now is time to map them to a variable that FAST search is able to understand.

The next screnshoot shows the CP description, and to which indexed filed will be index in FAST server.

This screenshot tell us the property belong to Sharepoint, The id of the property set for SharePoint columns this is: '00130329-0000-0130-c000-000000131346' (green square) , it is a basic type text (purple square), and there is a unique values (orange square).

You can get that infrmation also by using powershel cmdlet Get-FASTSearchMetadataCrawledProperty

Next step should be run the full crawl for the FAST Query search Service applicaton,  in order to get the map done between the two properties.

Until here the indexing part in this blog's series. the second part regars the seraching for the properties from the Qrserver ship on the FAST server.

NOTE: If you want to see how the source data get indexed as metadata in the indexer, see the next section enable the FFDumper . this operations must be done before the crawl

To enable your FFDumper debugging process,

1) go to FASTSEARCHetcconfig_dataDocumentProcessor and open theoptionalprocessing.xml file.

2) Here you will want to change <processor active=”no” name=”FFDDumper”/> to <processor active=”yes” name=”FFDDumper” />.

3) Reset your document processors by typing reset psctrl in your FAST PowerShell.

 In the c:\fastsearch\data\ffd\ *.* you will find the log that show how the column value get associated to the CP.

We could find also the values for ows_workphone, which will help us for the next part of the blog, the searching part.

The first way it is confirm the web application has associated the correct services.

So navigate to your Central administration || Manage service application || highlight the service application that is supposed to use the FAST search server || and click in the ribbon icon “properties”.

Scroll down the configuration page and check the content distributors’ field points the server’s name where FAST search server is installed; the default port should be 13391.

Now we know that at least the FAST server is involved in the crawling, which is a different topic than performance crawling or even a more basic question, if my crawler is sending data to the indexer?

A more advanced way is looking into the FAST server log.

The file c:\FASTsearch\var\log\procserver\ procserver.log should show you a similar entry like below, when you parse the file looking for the name of the Sharepoint server.

This entry will answer the question about if the Web Search Service application (FAST connector) is sending the data to the FAST server.

Other ways to troubleshoot you Web Search Service application (FAST connector) is my previous blog Troubleshooting the connection between the FAST Search server ContentDistributor and Sharepoint 2010

But it won’t help you in a Performance case.

Hi I followed this link http://blogs.msdn.com/b/spses/archive/2010/04/09/installation-and-configuration-of-fast-search-server-for-sharepoint-2010.aspx to install FAST Search server for Sharepoint 2010.

And now the FAST connector is not finishing the crawling. It is using many hours to crawl a little amount of document (~1000).

Well there is a chance the Sharepoint Crawler does not connect with the Content Distributor in FAST server.

In this blog I will explain:

How to ULSViewer  and netmon 3.4 to check up if your Sharepoint 2010 server connect with the Content Distributor in FAST, and what to do if there is no connection between them.

Step1:

Check up the Uls viewer.

Main while the FAST crawler is running open the uls viewer file> open from > ULS> select 14 folder >open

Configure the filter as simple as this, assuming you did not change the Content distributor port, by default 13391.

You should see the next error. Failed to connect to <fast server>: 13391 failed to initialize session.

Step 2:

Configuring Netmon to grab the information.

Open netmon in the Sharepoint server, before you run the capture, set up the filter as follow, where you aim the port 13391 (your customized CD port) and the FAST server ip address.

Then you should see like the next tcp trace where all the frames are sent but none is established.

There are several reasons for this behavior:

1)      The CD is stop so go to your FAST server and start up the Contentdistributor service.

2)    The connection to Content Distributor through SSL needs to be configured.

Log onto MOSS14 node as contoso\fast14

Open a SharePoint 2010 Server Management shell with the “Run as Administrator.

Copy the FASTSearchCert.pfx from the FS4SP node to the MOSS14 node:

copy \\FS14\c$\FASTSearch\data\data_security\cert\FASTSearchCert.pfx .

Copy the securefastsearchconnector.ps1 script to the MOSS14 node.

copy \\FS14\c$\FASTSearch\installer\scripts\securefastsearchconnector.ps1 .

Run the script with the certificate as input:

.\SecureFASTSearchConnector.ps1 –certPath FASTSearchCert.pfx –ssaName “FAST Connector” –username “Contoso\fast14”

You are prompted to enter the certificate password. Use the certificate password that you supplied while running the post-setup configuration of FAST Search Server 2010 for SharePoint

Wait for osearch14 to be restarted and connection validation to the Contentdistributor.

3)      The tcpchimeny and the task offload are not disables.

Running the next command in all FAST nodes might help:
netsh int ip set global taskoffload=disabled

netsh int tcp set global chimney=disabled

4)      You are installing the FAST Server in a box that contains any Sharepoint 2010 Server component.

 http://technet.microsoft.com/en-us/library/ff381239.aspx#BKMK_SoftwareRequirements

When the issue is fixed you should see the next netmon capture

As you can see there are a new SSL entries that confirm the communication have been established.

That’s all folks

Why to repeat a thing if it has been done properly….

But you still do not know if you are using the fast server, do not give up, I will try to explain how you could troubleshoot the problem.

In order to connect the previous blog and this there is a simple difference and it is the title of the created scope. In my case the name is fast and not vista like in the above blog .

Creating a site scope

To create a site scope, add the –OwningSiteUrl switch to your command line:

New-SPEnterpriseSearchQueryScope -SearchApplication "Search Service Application" -Name MySimpleScope -Description "My Simple Scope" -DisplayInAdminUI 1 -ExtendedSearchFilter ‘title:fast’ –OwningSiteUrl “http://myhost/sites/mysite/”

The first thing is checkup if the query reaches the FAST server.

So submit a query that uses the scope.

Log on to the fast server and open the path %FASTSEARCH%\var\log\querylogs.

Sort the “date modified” column by descendent and open the latest file, with your favorite text editor.

Tail the string qtf_parskw:filter.

If your search scope query hit the qrserver then you should get similar entry

As you can see the search scope query reach the server.

If your Search scope query does not return anything then is time to troubleshoot the filter

Let said that you used a little bit more complex filter and you don’t use the title and you use the path like this configuration.
New-SPEnterpriseSearchQueryScope -SearchApplication "FAST Query SSA" -Name "Documents Site" -Description "My Documents Library" -DisplayInAdminUI 1

$scope = Get-SPEnterpriseSearchQueryScope -SearchApplication "FAST Query SSA" -Identity "Documents Site"

$scope.Filter = 'or(path:starts-with("http://contoso.com/sites/it/en/docu/"), path:starts-with("http://contoso.com/sites/it/IT%20Documents"))'

$scope.Update()

So when you open the query logs everything seems right.

Well the filter hit the Qrserver but I don’t get any result .

Let see closely the querylog entry.

Why it is happen?

It is because you copy the url from the IE and paste in your script.

Using an url encoder utf-8

However the goal of this blog is not tell you the especific error you made, it is how to use querylog to troubleshoot the search scope problem