Best practices advice on scheduling

myITforum.com Community Forum

Home Forums Blogs Live Support chat Search Articles Wiki FAQ Email Lists Register Login My Profile Inbox Address Book My Subscription My Forums

Photo Gallery Member List Search Calendars FAQ Ticket List Log Out

All Forums RSS Feed Subscription:

Best practices advice on scheduling

View related threads: (in this forum | in all forums)

Logged in as: Guest

Printable Version

All Forums >> [Management Products] >> System Center Products >> System Center Configuration Manager >> Best practices advice on scheduling

Page: [1]

Message

<< Older Topic Newer Topic >>

Best practices advice on scheduling - 8/7/2008 12:07:26 AM

theIrish1

Posts: 7
Score: 0
Joined: 11/26/2007
Status: offline

Just wondering if anyone had some advice for best practices on the following 2 things:

1. Updating collections on a schedule. How often is TOO often? I have many collections that I would like to update fairly frequently... as in every 3-4 hours if not more often. I wouldn't think this would produce a deluge of database activity, nor a ton of sms activity as it is basically just re-running the membership queries on that schedule and updating the collection memberships in the database. (however, that is logically speaking, and I have been wrong before when trying to use, you guessed it, logic)

2. Hardware inventory schedule. How often is TOO often? I use hardware inventory not only to pick up actual hardware information about PC's, but also to pick up the NOIDMIF files produced by 10-12 audit packages that I run on a daily basis. I would like to pick up this hardware information on a frequent schedule as well, say every 3-4 hours. I have had problems in the past where SMS will abandon old MachineID's and produce new ones for the same machines. This creates duplicate records in the database, and therefore in the reports I run as well. These old abandoned MachineID's hang around until they are far enough out of date(old enough) that the site maintenance jobs purge them. Seeing dup's in reports creates doubt, and increases the overhead as far as time spent by me scrubbing the reports when one is requested. Can anyone think of a reason for this activity other than doing the inventory to often?

Anyway, any kind of advice or insite that anyone can offer is appreciated.

Thanks,
Irish

Post #: 1

RE: Best practices advice on scheduling - 8/7/2008 3:07:43 AM

inbaraj.suresh

Posts: 4
Score: 0
Joined: 8/5/2008
Status: offline

(in reply to theIrish1)

Post #: 2

RE: Best practices advice on scheduling - 8/7/2008 3:08:42 AM

inbaraj.suresh

Posts: 4
Score: 0
Joined: 8/5/2008
Status: offline

This article will provide you with the steps need to follow hardware inventory from the local PC into the SMS / ConfigMgr database. This will give you an idea as to where any breakdown occurred during the hardware inventory flow if there is an issue. Although you can use any log file read you like, I recommend using SMS Trace.

Starting on the Client, start the Hardware inventory

Open the Inventory.log using SMS Trace. Confirm that the Hardware inventory has started

Wait for the inventory to complete.

If you get to this point then you know that the PC has transferred its inventory to the MP and therefore is not a problem on the client. If you don’t get the above line then the problem is on the client.
Move to the MP server and open the mp_hinv.log

Now you can see that the Hardware inventory from my client was received by the MP and moved into MP outboxes.
Now move to your site server; in my case it is the same server. Open the mpfdm.log. Notice that the file is moved from the MP outbox to authenticated dataldr.box and that the file name changes.

Open up dataldr.log
Notice that the file is moved into the dataldr.box\process directory and then it is renamed to X??????????.mif

And finally you can see 687 stored procedures were executed and data is now within the db.

And Resource Explorer shows that the data has been updated.

(in reply to inbaraj.suresh)

Post #: 3

RE: Best practices advice on scheduling - 8/7/2008 7:03:14 AM

skissinger

Posts: 2114
Score: 134
Joined: 9/13/2001
From: Sherry Kissinger
Status: offline

As with anything "It depends" (don't you hate that answer!). For #1, It depends on the # of collections you have, the frequency, and your hardware. Is it every collection you want to update that frequently, or a subset? Monitor your site status & log files before you make any change, then modify (and keep track of) the collections you care about the most, and monitor your site status & site server log files for a few days to a week. If the increase doesn't create a backlog, you know it's OK.

For #2, outside of a lab environment with <10 computers with less than daily; the most frequent I've heard of is daily HW Inventory. Again, it'll depend upon your environment. Once you've finished testing #1 (and I wouldn't recommend changing both simultaneously), then do similar tests of a before and after of changing your HW Inventory frequency.

Dups in the database--I get those when a computer is reimaged. That's because we don't transfer the GUID in our reimage process. Other than duplicate GUID issues when it makes sense for a computer to get a new guid, there's usually a reason for dups in your database.

_____________________________

mofmaster@smsexpert.com (version 2007) | http://www.smsexpert.com | http://www.sccmexpert.com
My Blog
Microsoft MVP - ConfigMgr

(in reply to inbaraj.suresh)

Post #: 4