Full Crawl
Full Crawl of Content source will re index all the content
from beginning
Important Points to Consider:
- If new managed property has been introduced, we need to run Full Crawl of content source
- If new crawl rules are created/updated/deleted, Full crawl of content source is required
- If incremental crawl has been failed
- If software update or service pack has been installed on the servers
- Expensive in terms of performance issues
Incremental Crawl
Incremental Crawl of Content source will only process those
items which are changed since the last crawl happened.
Important Points to Consider:
- Most preferred after the full crawl has been done.
- Does not hamper the performance as it will crawl only modified documents not the entire content source.
- The incremental crawl will retry items and postpone crawling the item if the error persists.
A limitation with the Full Crawl and Incremental Crawl is we
cannot schedule both to execute parallel. For example if the Full Crawl is
already running then the Incremental Crawl cannot be triggered until the Full
Crawl completes, if you try to Stop Full Crawl then also it is mandatory to
finish at least once successful Full crawl before triggering any Incremental
Crawls.
So Microsoft has come up with the concept of Continuous
Crawl
Continuous Crawl
With Continuous Crawl you can maintain the content index as
fresh as possible.
More than one continuous
crawl can run in parallel
one deep change
will not result in degraded freshness on all following changes
The impact of a "Continuous Crawl" is the same as
an incremental crawl.
At the parallel execution of crawls, the "Continuous
Crawl" within the parameters defined in the "Crawler Impact
Rule" which controls the maximum number of requests that can be executed
by the server (default 8).
