Indexing a content database then replacing it with a more up to date version of the database is full crawl needed

Question

Very hard to define a title for this question.

I have a Test farm. I have a copy of the content database from production. We have indexd the farm for search. Every 3 months or so we get a new copy of the content database from production to test against. After I have switched in the new content database would I need to perform a full crawl again or would an incremental crawl work?

Answer 1

Follow the instructions on MDSN Move content databases

When you read the guidlines plan for crawling and federation it says that

The system does a full crawl even when an incremental crawl is requested under the following circumstances:

• A search administrator stopped the previous crawl.
• A content database was restored, or a farm administrator has detached and reattached a content database.
• A full crawl of the site has never been done from this Search service application.
• The change log does not contain entries for the addresses that are being crawled. Without entries in the change log for the items being crawled, incremental crawls cannot occur.

Answer 2

I advice to have a full crawl. Especially if you have Metadata Service used. As far as I understood you do not copy anything except for the content DB (no Metadata Service DB copied) so the ids of the tags can differ from environment to environment and they are 100% different if created by hand. This is only one example.

And of cource the full crawl is preferable if the producrion has the content that is missing in the test environment. With no full crawl the content that has never been crawled will not be crawled incrementally.

Answer 3

This question should help you to migrate your production to your test farm. Reconfiguring SharePoint 2010 Portal An incremental won't suffice, I must use a full crawl after the migration in this instance, the crawl log will also show up errors in the replication.

Since the database content has changed between the updates it is a must in this case. You can get away with an incremental crawl when doing exact replication, but this is not one of those cases which it appeared to be in the first place.