Hotfix for the SP2 issue is available now

Service Pack 2 Update.

The public update for the Service Pack 2 expiration date issue<http://blogs.msdn.com/sharepoint/archive/2009/05/21/attention-important-information-on-service-pack-2.aspx> is now available for download.

Whoo-hoo, I hope that you caught this one!

The update can be applied before or after Service Pack 2 installation. If the update is applied prior to installing Service Pack 2 it will prevent the expiration date from being improperly activated during installation of Service Pack 2, if it is applied after Service Pack 2, it removes the expiration date incorrectly set during installation of Service Pack 2.

Installation instructions and download links for x86 and x64 are available in this KB: http://support.microsoft.com/kb/971620

For 32 is this link: http://download.microsoft.com/download/2/F/5/2F51AB71-1325-49D2-9CB9-18DEC4780E99/office2007-kb971620-fullfile-x86-glb.exe

For 64, this one: http://download.microsoft.com/download/5/B/B/5BBD34A9-C528-42B0-8A5F-9A8997B25C32/office2007-kb971620-fullfile-x64-glb.exe

Profile properties import through BDC connection to Oracle SAP

Interesting fact was uncovered during our enterprise deployment project, and I wanted to share it.

The client has Oracle SAP that they are using across the globe, and we were supplementing user profile properties that have very limited information in AD through BDC to Oracle. Import through BDC was very slow, I mean VERY slow.

If AD full import was taking 1 1/2 hour to import 70,000 user profiles, then BDC part would take around 20+ hours. with this numbers, there was no way for us to meet SLAs.

After Oracle team, Microsoft team, and us started investigating the issue, some interesting things were uncovered.

Apparently in our ADF we were using simple select statement, to oracle side our “identifier” or record ID (ex:“win_no”) field was being passed from SharePoint as a numeric field. In the world I know, this is how you treat record ID.

But on the Oracle side this field was defined as a character. In this scenario, Oracle disabled the use of the index because the field types do not match. With the index disabled the process of retrieving records was slow.

The issue was solved by the “to_char” we put in select statement to make sure that BDC “does not lose anything in the translation”. Final testing showed that BDC import was brought down from 20+ hours to under 2.

When I was at NY User Group meeting in June, I have heard someone saying that their BDC import was taking about 8 hours to finish, and for a while, I thought that it was normal. Well… It is, if your field types do not match! :-)

Enjoy

Converting Hyper-v VHD to VMware

Real bummer. I had to power off my server that was running Hyper-v and the whole bunch of VMs. Until I get the hyper-v host back, I decided to continue my development on one of the machines by converting it into vmware. The general VMware converter that allows you to convert from VHD is not working with POWERED OFF hyper-v VHD. apparently you can convert it only as a physical machine by giving the IP address, which in my case is not possible as I do not have another Hyper-v host. I guess I have to be patient and wait for the server to get back

:-(

Publishing Search Results in MOSS

Never noticed it before, but after checking it on several installations of MOSS, it appears that the search is “ignoring” “start Date” on publishing pages and includes all documents that had not been published yet.

If you create for example an article and set publish date one week from now, you have to modify your search not to include these articles into search results until this date had been reached.

New York SharePoint User Group meeting June 3, 2009

It is a short notice, but I urge you to attend this meeting.

Jim Kane will present on "Challenges and SharePoint Enterprise Deployments (& a few solutions)"

This presentation will focus on the many and varied challenges of implementing SharePoint on an enterprise level, whether for a small company or for a large company. Enterprise deployments present unique issues; from expectations through politics, from infrastructure decisions through going out-of-the-box or customizing. Deploying SharePoint across an enterprise can get messy quickly once more than a few people are involved in decision making.
We'll look at some real-life examples of challenges, and some ways to meet those challenges. Be prepared to participate in the discussion!

I’ve known Jim for quite a while and the first time I met him it happened at one of the New York User Group meetings. Jim and I are currently working on the same project (enterprise deployment), but from different side. Sides do not really matter form the project point of view, but I’m curious to find out how the challenges that we all are facing are being viewed from a different perspective.

I’ll see you there :-)

FAST ESP 5.3 Installation trick

The installation is not that cumbersome, but the trick here is to install the right prerequisites especially java JDK(Java SE Development Kit with JavaFX (JDK 6u13 / FX 1.1)), you can find it at http://java.sun.com

If you do not have the right version of JDK, your installation will fail on Windows ( never done it in other OS), unfortunately you can’t really find it in documentation, so you can spend sometime struggling with it.

One of the great features of the FAST installer is that once installation is done, FAST generates installation profile XML document that can be reused in the future to perform that same type of installation scenario, this is especially useful in multi-node scenarios.

FAST search engine for SharePoint PART 3 (Content Sources and Connectors)

Each “content source” is represented as a “Collection” within FAST ESP. Data is being fed into Document Processing Pipelines for refinement through the use of “Connectors” that are defined for a specific collection.

There are three types of connectors for FAST search engine:

1. FAST OOTB connectors
2. Third Party (proprietary)  connectors
3. Custom connectors, using FAST API

YES! FAST allows you to go against APIs

FAST OOTB connectors

Enterprise Crawler: used to feed content from Web Pages. Content sources, and many other settings for this connector are easily configurable through the Admin UI, including: Content Request rate, Start URIs, Include and exclude host name filters, content crawl interval, etc. Enterprise Crawler allows you to crawl unlimited number of start URIs, detects deleted content, and removes it from index, and retrieves both: static and dynamic content.

ESP File Traverser: traverses and submits files from file system to content pipelines in batches via Content API. Files that this connector serves can be in any binary or text format, as long as this format can be handled by processing pipeline (PDFs, TXT, XML, DOC, and many more).

JDBC Connector or FAST Smart Connector for JDBC: This connector uses database data or structured data for feeding into pipelines (Oracle, SQL, MySQL, DB2, etc). This connector uses JDBC driver that must be registered on the server prior to establishing a connection to the database. It extracts data to be indexed on a column level through the use of SQL query that you supply, this connector is managed through command line and through Web Interface as well.

Connectors mentioned above support content modification detection through the use of checksums as well as timestamps that are kept either in FAST built-in db or some other MySQL db.

Third Party Connectors 

There is not much I can say about these connectors except to list some of them: Lotus Notes, WebSphere, Exchange, Documentum, Hummingbird, and of course SHAREPOINT. So even if you are not looking to upgrade to the SharePoint 2010 when it becomes available with FAST, you still can integrate with SharePoint without reinventing the wheel :-)

Enjoy :-)

Technorati Tags: FAST,FAST connectors,search,database search,web pages search,sharepoint