A Refresh
walk behaves like a Refresh All
walk, but it doesn't
check every URL every walk. The Parametric Search Appliance pays attention to how often each
URL changes, and schedules checking the URL less often if a URL isn't changed.
When a Refresh
walk starts, it only refreshes URLs that are scheduled for
update at the start of the walk.
The idea is that if a profile is doing nightly walks and a URL hasn't changed in the last 6 months, it probably doesn't need checked EVERY night. It can be checked every 2nd night, every 3rd night, every 5th night, etc. as it continues to not change.
Refresh
walks - Refresh
walks are useful
with a large (200k+ URL) collection of content that doesn't change very often,
where the collection is too large to perform a Refresh All
walk in a
timely manner and dataload isn't possible. Refresh
walks can finish
much faster than a Refresh All
walk. This allows another walk to start
sooner and frequently-changing content to be re-checked sooner, instead of
taking the time to finish refreshing all of the almost-never-changing content
first.
The downside of Refresh
walks is that if a URL whose content rarely
changes does change, it may not be picked up in the next walk because
that URL may not be scheduled to be checked in the next walk. It may be
worthwhile to schedule or manually launch an occasional Refresh All
walk to check content slightly more often.