"Hank Arnold" <> wrote in message
news:...
> Wow!! Glad to see it finished, but as others have said, that's **WAY** too
> long. Either you have "real time" scanning enabled or there are hard drive
> problems. I would run diagnostics on the hard drive to make sure...
Actually.. several hours is not unlikely at all if:
* It has never been run before and the server has been operational for some
time
* Definition updates are synchronized
* Definition updates are auto-approved
* Superceded updates are still approved, or have never been manually
declined
* The first option is run by itself without first having run the other
options.
My recommendation for using the Server Cleanup Wizard (particularly for a
first time execution on a veteran machine) is a multi-pass methodology as
follows:
1. Delete all stale computer entries. (Option #2)
When analyzing updates for computers that have reported those updates as
needed, each computer entry in the computer table must be joined to every
update to analyze status. Removing dead computer entries from the list
significantly reduces the number of joins that must be made for those
analyses.
2. Expired and Superseded updates. (Options #4 and #5)
This pass does not require the removal of any data, merely the flip of the
Declined flag on the update definition. It generally runs very quickly. The
"expired" updates option is based on a single flag value; the "superceded"
updates option has to be joined with computers and groups (for approvals),
but only involves a subset of the updates table (those with
isSuperceded=True).
3. Unused updates and update revisions. (Option #1)
This pass actually deletes metadata from the database. It's the deleting of
data that is time consuming because indexes must be rebuilt/reorganized
anytime data is deleted from a database. Note, however, that the only
updates that are deleted are those that are expired, or those that are older
revisions. In addition, the search for these updates is somewhat resource
intensive because it involves joins to computers and group (for approvals).
The first part of the pass is only a subset (isExpired), but the second part
of the pass has to evaluate every update for the presence of revisions and
evaluate the approval status on each of those revisions.
It's possible that reorganizing the indexes prior to running this step
(using the Database Maintenance script documented in the Ops Guide) can
speed it up; however, in the end, you'll still likely invest the same amount
of time on the entire process.
4. Unneeded update files. (Option #3)
After all of the metadata has been properly classified (Declined or
Deleted), then it's appropriate to clear out the filesystem.
BIG NOTE: If you've enabled auto-approval rules for Definition Updates, you
should be running the Server Cleanup Wizard on a *weekly* basis. Forefront
Client Security updates generate 10-12gb of file content per week, all of
which has an average lifespan of about 6 hours. By running the SCW on a
weekly basis, you can keep the FCS definition files down to about four
generations of files, and significantly reduce your overall file storage
consumption.
--
Lawrence Garvin, M.S., MCITP:EA, MCDBA, MCSA
Principal/CTO, Onsite Technology Solutions, Houston, Texas
Microsoft MVP - Software Distribution (2005-2010)
My Blog:
http://onsitechsolutions.spaces.live.com
Microsoft WSUS Website:
http://www.microsoft.com/wsus
My MVP Profile:
http://mvp.support.microsoft.com/pro...awrence.Garvin