Skip to main content

ATG search installation and search engine log files

Here I would like to explain how your ATG search folder structure looks like after you do the ATG search installation and also about the search engine log files.

To get the ATG search, you need to install the ATGSearch<version>Windows.exe after you do the ATG installation. When you install ATG search, the generated folders include an OS dependant folder. In the below screen shot, you could see the "i686-win32-vc71" folder and this is generated since the ATG search ins installed in a windows machine.


The ATG search engine is actually an exe file inside this "i686-win32-vc71" folder. Below screen shot give you an idea about the "atgsearch.exe", which is the search engine application. This search engine application (atgsearch.exe) is started by a remote launcher call from the BCC search administration UI. 



The log files like "atgsearch_20120106112413_568.log" are the search engine log files where you could see the engine logs like the index file used by the search engine,enigne load ,enigne throughput and query response time,.. You could enable the debug logging for the search engine to find the different request processing by the engine including the ip address of the requesting host. Also the search engine start and end time can be obtained from this log.



The engine log will show slow request based on the <SlowRequestThreshold> parameter in the AEConfig.xml, like <SlowRequestThreshold>5000</SlowRequestThreshold>. If you set engineLogLevel to D in launchingService component, you will be able to see the ip address of  of the requesting host.

                 engineLogLevel = D

Example of the search engine log with requesting host ip:
2011/03/22 11:12:59 : DEBUG : Server : 0 : Accept socket 12068 at port 4790 from IP 100.222.46.124
2011/03/22 11:12:59 : DEBUG : Server : 0 : AEXmlWorker::Shutdown
2011/03/22 11:12:59 : INFO : Server : 0 : Shutdown received
2011/03/22 11:12:59 : DEBUG : Server : 0 : Accept socket 12068 at port 4791 from IP 100.222.46.124
2011/03/22 11:12:59 : INFO : Server : 0 : Waiting for pending requests to complete.
2011/03/22 11:12:59 : INFO : Server : 0 : Exiting engine
2011/03/22 11:12:59 : INFO : Server : 0 : Scheduled shutdown.


Consider the following search engine log :

2011/03/11 09:36:35 : INFO : Server : 0 : load: 0.6875 throughput: 1.72804 q/s time: 6.60434 s

Load  of .6 means that 60% of the search engine thread are active. Throughput is the number of requests that the engine is able to server per second. Time is the average response time.

So load is specified as percentage and 0.6875 is roughly 69 percent of load, where as anything greater then 1 would suggest the search engine is overburden. Throughput is telling us how many requests the engines can handle across all the cores of the engine hosted box. So throughput is kind of like saying how many request can be placed into the bucket. But it doesn't say how fast the objects in the bucket are moving.



Comments

  1. Hi,

    I have a problem with Service Center's search engines, ATGOrder and ATGProfile. Both has a scheduled shutdown at midday and midnight causing searches and live indexing to stop. How can I turn the scheduled shutdown off?

    Thanks in advance and good blog!

    ReplyDelete
  2. you mean search engine for atgorder or for atgcatalogue ? U mean at midnight ur search engine s are going down ? Any scheduled search indexing is configured for the search engine ?

    ReplyDelete

Post a Comment

Popular posts from this blog

Google Chrome shortcut keys

If you are a Google Chromey guy, please find below the list of shortcut keys for some of the most used features  :-) Find more shortcut keys @  http://www.google.com/support/chrome/bin/static.py?page=guide.cs&guide=25799&topic=28650

Apache Lucene quick links

Lucene home page –> http://lucene.apache.org/ Download Lucene from –> http://www.apache.org/dyn/closer.cgi/lucene/java/ Lucene API Doc –> http://lucene.apache.org/java/docs/api/index.html Lucene docs for each release –> http://lucene.apache.org/java/docs/ Where can I get help from –> java-user@lucene.apache.org Lucene wiki  –> http://wiki.apache.org/lucene-java/FrontPage Lucene , how to improve search speed –> http://wiki.apache.org/lucene-java/ImproveSearchingSpeed Lucene , how to improve index speed  –>   http://wiki.apache.org/lucene-java/ImproveIndexingSpeed Lucene FAQ –> http://wiki.apache.org/lucene-java/LuceneFAQ

How to convert your Blogger Blog to PDF ?

You can use a website called "blogbooker" @  http://www.blogbooker.com/blogger.php   to convert your Blogger Blog to a PDF . Please find the steps below : 1. Save your blog as an xml using Blogger Settings - Other - Export Blog option 2. Go to the website " http://www.blogbooker.com/blogger.php " and select this XML , give your blog address and select the options like date range, page size, font, ... 3. Click the  "Create Your BlogBook" button to view and save your blog as PDF

Basic design decisions for a commerce search setup ( with an ATG Search view)

In this blog I would like to explain the basic set of configuration/design decisions needed to setup an ATG search project. Most of these design decisions are common for all Enterprise search applications. 1. Decide the searchable properties :   This means the properties that the business want the user to search in the ecommerce platform. In ATG search these are configured as the text properties in the product-catalog-output-config.xml ( the definitionFile of the \atg\commerce\search\ProductCatalogOutputConfig). Usually the displayName of product/sku, displayName of department/category/sub-category, skuId, brandName are the properties configured as searchable. 2. Decide the search refinement properties or the faceted properties :   After a user search for a keyword, search refinement is the next step done to filter his results. ATG supports the search refinement using the Faceted Search concept. Read more about facted search @...

ATG Order update - InvalidVersionException and ConcurrentUpdateException

ATG repository item descriptor can have the version property. The atg.adapter.gsa.ItemTransactionState holds this version information. For example consider the Order item-descriptor. It has the version property defined against the table dcspp_order. Means, the dcspp_order table has the column version which defines which version of order is currently in the DB. Each order update flow will update this column.  <property name="version" display-name-resource="version" data-type="int" queryable="true" readable="true" column-name="version" hidden="false" category-resource="categoryInfo" expert="true" required="false" cache-mode="inherit" writable="true">     <attribute name="uiwritable" value="false"/>     <attribute name="propertySortPriority" value="30"/>   </property> ------------------------------- Du...