MMS 2008 - update
Summary: now that we're home, it's time for general observations and other follow-up. Actually, I've been meaning to do that for at least a week now, but it's been busy times around here (and the weather has finally been decent).
The Conference Bag
On the myITforum WIKI, I have maintained a history of SMS/ConfigMgr bags. I have to figure out my account/password for that one, but in the meantime, here's a picture:
There's actually a lot of pictures of the bag out there. Someone even does a complete 'unboxing': http://www.myitforum.com/absolutevc/?v=910
I'd say this year's event bag was unique - it's clearly a backpack bag. Only one past bag had shoulder straps, and they were optional (the bag could also be used as a suitcase). But as a mountaineer I take backpacks seriously and the best I can say about this one is that it's a modern traveler's bag. It has no provision for a laptop, is too large for business meetings, and has no lap belt for serious loads, but does have provisions for devices. Most people did carry them around throughout the event, so I suppose they had some utility, but I'm not sure of the ongoing value. The intel logo is prominent, so I suppose the sponsor is happy.
Technical Updates
In my "Going Big" presentation I mentioned that when going to ConfigMgr one thing you should watch out is 'rogue' WSUS installations in your environment that will conflict with SUM (or which SUM will conflict with). Fortunately there's an easy way to avoid the grief - get your current SMS infrastructure to look for them. That's done quite easily with a hardware inventory extension for WSUS. We do that by collecting this registry key: HKLM\SOFTWARE\Policies\Microsoft\Windows\WindowsUpdate\WUServer
On the client health front, there have been a few updates (which I will also incorporate into the relevant posts):
-
Greg Ramsey has shared their
ConfigMgr Client Troubleshooter. I've used it a bit and it has some handy features. It has a bit of 'many clients' functionality but it's not meant to be an enterprise-scale solution. Manual inspection and repair is always going to be a part of any client health strategy, and this tool really helps in those situations. So it is a good addition to your client health toolkit.
-
- Chris Stauffer has begun an ambitous effort to produce a client-side client health checking script. It's a script that could be run as a scheduled task, or as a regular advertisement (for partially healthy clients). There's a lot of functionality to be packed into a vbscript, but I know Chris can do it, especially with some help from the community.
-
In case you missed it, John Nelson
added a great comment to one of my earlier postings about WMI reliability hotfixes. In particular, on how to verify they've been applied to your clients. I haven't tried his idea yet, but it looks like a good one to me.
-
One suggestion I heard was for a Wiki on client health. That way we can work together to fill out the client health story, and provide the "guidance" that I often talk about to each other. Fortunately, myITforum has a wiki:
http://www.myitforum.com/myITWiki/. I'm trying to figure out how to make it work.
-
Dan Thomson and I had a great conversation about many client health topics, and one of his suggestions was that a computer startup script could also be set up to run as a scheduled task on each client. That would mitigate the main limitation of computer startup scripts, which is that they only run as often as users reboot (especially while on the corporate network). Somone else said they also send there's out as regular software distribution. That only helps with clients that are healthy enough to run a software distribution, but since the content and policy can be cached, that's actually more clients than you might expect. I'm embarrassed to say I didn't have either of these ideas years ago
-
One organization I've talked to recently has developed a 'control' that sits on each of their clients and reports (via hardware inventory) the response times of their DPs and MPs. I'm hoping to encourage them to share that with us.
-
A manager that I talked to at the conference said that there was an upside to his staff doing a lot of manual client health remediation: they come to understand the clients well and get really good at client-side troubleshooting. He even observed a significant improvement in morale - techies like tricky technical challenges!
-
Of course I see a lot of specific client health issues in the course of my work, or in the course of client health discussions in the community. WMI issues may be the most classic example. But in the course of conference convesations (or conference-inspired conversations), I've heard some new issues:
-
Disk corruption: if the disk is having problems then BITS may fail to download content, or there may be other kinds of client health issues. Running chkdsk a few times, even without the switch for fixing the problems, often helps. The customer that reported this one sees this daily.
-
Another customer reports the startup details for BITS were disappearing on many of his computers. We did a little poll of other customers and none had seen this, but maybe they weren’t looking.
-
Someone suggested that one measure of a healthy client is one that has a current patch mgmt catalog, since patching is a core computer management service and many parts have to work well for the catalog to be up-to-date. So it makes sense to watch for old catalogs.
-
One idea that should be easy to verify, but that I haven't had time to do just yet, is that if the PATH is messed up then ccmexec won’t start. For example, if there is a simicolon in an inappropriate place, or two backslashes when there should be one.
- Another suggestion for verifying a client is healthy is that the SMS/ConfigMgr WMI namespace has classes for recent content and policy downloads, and if those are old, incomplete, or have errors then the client is unhealthy
-
Running out of disk space (anything less than 90 meg) is an obvious way to cause client health problems, but what can you do to automatically but safely fix that problem? Deleting the restore details for old patches, and clearing the SMS/ConfigMgr cache of old content are good suggestions
-
One customer observed that clients with a missing RSOP namespace seem to cause problems, but this problem also easy to fix (recompile the MOF). I haven't verified the details on this one, but I do recall seeing a the RSOP namespace
- An up-to-date SMS_Def.mof version would also be a good indicator of a health client
-
Another way for WINS or DNS to have issues is for DHCP to be set up incorrectly. Watching the details at the client side would help to identify these issues
General Observations
The unfortunate reality is that many of the most interesting things I get to do at the conference are not bloggable. Not things that would change the world, but discussions with people that expect some discretion, personal discussions, or discussions that could be easily misunderstood when taken out of context. I suppose the point is that there is a lot of 'backchannel' activity that happens at the conferences. And even if you're out of those loops, you're actually involved, in that your comments clarify positions and give credibility to the whole market/community. I know I'm out of the serious loops, but there's a lot of those backchannel conversations at the conference. So there's a lot I've been thinking about since the conference but I can't get into much of it here.
Overall I can say that this was one of my best MMS conferences. I attended fewer sessions than ever before - in fact only one (I'll listen to many more on my Zune later) . But the conversations were non-stop and highly varied. Every year I feel more "at home" at the conference. The food-for-thought was nonstop.
At a personal level, I can say that I feel energized in a way that I haven't felt in about a year. It's like youth has returned. I have to think that has to do with the quality of the conversations. But maybe the quality accomodations, good food, nice trip, etc. helped. In any case, I am inspired. Now to get back to work!...