Search Service Application : old entries to remove to be accurate

Since last week, I've nearly done all to adapt the Search Crawl of our corporate mySite, so it'll match the ne requirements.

During previous tests last November, some old entries were not able to be removed.

I've set new filters, exceptions, etc …

All done, and always executing a full crawl after each changes.

Then, we had to enable the "NetBiosNamesEnabled" option, as our domain is known in 2 different names.

All was redone and working, except for the crawl.

I was expecting that a DB was keeping trace of the old entries, and didn't clean up that database.

A kind of index.

With the help of a consultant, I've explained with as much details as possible what I've done, and what I'm faced with.

Today, he explained that the Indexer in the Search Service Application had to be reset and the mySite contentdb still contains the wrong "domain" users, users that had been crawled at the time before enabling "NetBiasNamesEnabled".

The only way to get this out of the “system” (read: Indexer) is to reset the indexer.

What is the impact of resetting the indexer?

It depends on how big the indexer files are, how long it will take to build up a new indexer file.

Because with resetting the indexer, the existing indexer file will be deleted, and a new indexer file will be created. It will be empty.

This means for the end user that they cannot perform any search anymore until the indexer file has been build up again.

How to build new indexer file after resetting the indexer?

By running a Full Crawl. And this can take some time, depending on the quantity items that have to be crawled.

So the action plan is :

Reset the index

Do a full crawl

Then, after testing, I saw all the wrong entries removed from the search in mySite.

That's all folks