Business Desktop Deployment 2007 RTM

Microsoft has announced the general release of BDD 2007.  This free download is a kit aimed to aid administrators and consultants in the deployment of Windows Vista and Office 2007.  Microsoft says:

Business Desktop Deployment (BDD) 2007 is THE best practice set of comprehensive guidance and tools from Microsoft to optimally deploy Windows VistaT and the 2007 Microsoft® Office system. BDD 2007 has undergone many changes to significantly ease the deployment of Windows Vista, including new out-of-the-box imaging technologies, XML-based migration scripts, new tools for image engineering (ImageX and System Image Manager), and new unattend XML format. Another huge breakthrough with Windows Vista is Hardware Abstraction Layer (HAL) independence and language neutrality; this means that organizations can get to a minimal number of standard images across multiple hardware types.

Some of the key information for using the Windwos PE solution to deploy Vista is scattered all over the documentation.  I’ve some blog entries on my site with stpep-by-step’s on doing the basics.

Using MOM 2005 for Service Level Reporting

I am a big fan of Microsoft Operations Manager 2005. I’d been using it since I deployed the beta in a limited basis on a production network to monitor critical servers that had been unreliable. I’d previous experience with Computer Associates TNG 2.2 and 2.4 (I was a certified consultant) and with Hewlett Packard’s OpenView (an extensive trial program). I wasn’t prepared for how extensively MOM monitored and reported on my network. From the moment it slapped my wrists for badly configuring the MOM Operations database and telling me how to repair it I was convinced that this was the only product for me. We ran a fair tender process but no other contender could come close to the functionality of MOM.

One of the big selling points is MOM’s ability to collect data on and alert you about server and application performance and reliability. This data is archived in the reporting database and you can run extensive reports. Some examples are:

 

  • We administrators know that we should run baseline performance reports and run comparisons over extensive periods. We never do it because we never have time and we don’t know where to store this data. Heck, we often don’t know what to monitor because you can’t be 100% sure of what will be a future bottleneck, mainly because you are the last to know how the business plans its future growth. MOM makes this easy … it collects all key metrics and stores them so you can run a reports on past performance and compare it to current performance. This allows you to scale future growth.
  • Application stability and server fault information is painful to track by hand. Sure, we can use electronic tools such as InfoPath and SharePoint to keep electronic forms but we still have to manually enter the information. MOM keeps track of all faults. These can be compiled in reports that are generated on a regular basis.
  • Alert resolution times should be monitored and tracked. Again, this can be done by paper or electronically. If you are using MOM 2005 properly, your administrators should be resolving alerts in the Operator Console and logging the cause and resolution of the problem in the are provided in the Company Information section in the alert, thus keeping a history and allowing administrators easily resolve similar problems when the arise in the future.

You can probably see where I’m going with all of this. Lots of organisations need to be able to produce reams of data for Service Level Agreement (SLA) reporting. Gathering this data could be really difficult but thanks to MOM and the reporting database, it’s easy to access this data. The difficulty is, how to I gather this data into a single report?

MOM reporting is great at producing small chunks of similar data in reports. But a report that gives you all relevant performance, fault and resolution data for all servers and applications does not exist. This means there is no single report for producing the information that you need to do an SLA monthly or yearly report. I looked into how this could be done. I can do some basic SQL and I understand views, joins, etc. But what you find under the covers is … well … it isn’t easy. My conclusion was that to use the tools provided in Visual Studio to customise a single report or maybe just a few reports to do all of this could take a team of SQL and SQL Reporting Services experts a long, long time.

This was a problem for me. My client required a solution to provide monthly reports based on data collected by MOM 2005. Ideally this solution had to be as automated as possible and present data in a meaningful manner that non-technical people could access.

I came up with a solution that works, is completely automated and can be presented as a document or as a web service. I’ve documented the solution here and I hope it will be of some use to somebody out there.

Generating the Data

There is a 99% chance that the data we need to access in a Microsoft centric network is available in the MOM reporting database via the MOM reporting website. You can access this via http://<ReportingServerName/Reports. Some administrators may have used SSL so that may end up being https instead of http. You’ll also need access to the reporting website.

My first step was to indentify the reports and metrics that I needed. I documented these. I’ve found over time that Operational Health – Performance Analysis was my primary reports for performance metrics, rather than one of the performance reports under the OS reports.

My next decision was when to run these reports. Remember that you’ll be running one report of each kind per server. If you have 5 reports and 100 servers then this is going to be 500 reports. This can eat up CPU cycles. If you do have 100 servers and reporting is important to you then a dedicated reporting server is something you should consider, i.e. a 2 server deployment of MOM 2005. If you are running 5 types of report then I would recommend that you set up 5 different shared schedules in MOM Reporting. A shared schedule enables you to define your define your schedule once and then reuse it. This save a lot of mouse clicks. I’d place them fairly close together. How close? That depends on how long each one will take to run.

Now we need to go back to our reports. We’ll be working with linked reports which are related clones of reports that use an existing report definition rather than a new one. We can create some defaults in these linked reports to save on mouse clicks. They’ll appear to an end user as normal reports. We’ll need somewhere to locate them. In MOM Reporting, create a new folder under the MOM 2005 reports and rename it as “Custom” or something like that.

Now you should navigate to each report that you want to use. Create a linked report using this template and store it under your new “Custom” folder. For example, I create a few linked reports based on “Performance Analysis”. Some would be use for processor metrics and some would be used for logical disk metrics. My linked reports were named appropriately.

Now I entered the properties of each report and customised the parameters based on entries that would always be entered into a report when it would be run. The aim is to minimised mouse clicks. For example, in my CPU Percentages report I defined counter #1 as Processor-%Utilisation-_Total and Processor-%Interrupt Time-_Total. Now, every time I need to run this report for each server (remember 100 servers by 1 report = 100 reports) I know that I’ve saved a bunch of mouse clicks by defining what is to appear in the report and when it is to be run (the shared schedule).

We’ll be generating a lot of reports as you’ve probably figured out. We’ll need to store them somewhere. On your MOM Reporting Server, create a file share to store the data. I created a hidden file share called MOM Reports$ and then created a sub folder which was named after each server that I would run a report against.

Once all of my linked reports are done we have to figure out how to automatically run them for each server. This can be done using a subscription for each report and for each server. This means that you create a subscription for each server for each report. My process here was to create a subscription, specify a shared schedule to utilise and store the resulting data as a TIF image file on the newly created file share in the appropriate server sub folder. Note that I configured my subscriptions to replace the previous exports with the new ones. Some subscriptions would require extra configuration, e.g. Logical Disk-% Free Space will require that you list local drives other than the default of C: for complete reporting. You’ll need to create a subscription for each server and for each report. That’s a lot of work but the saving grace is that you only have to d
o this once. After this your data is stored in the file share on a regular basis as a set of TIF images that are organised into folders that are named after the servers being reported on.

We’ve now generated our information. It’s hardly in a format that we can present to management or clients.

Presenting the Data

The aim of my solution is to provide two means of presenting the data in the SLA report:

  1. Information should be accessible via a web site or page.
  2. It should be possible to generate a document based on the information.

My solution works as follows:

 

  • I gather all of the data into a webpage or website.
  • Those who are responsible for reporting operations can print the website/webpage off as .PDF or use Internet Explorer to save it as a single web file archive (.MHT). Either of these file types can be presented to management or to a client and opened to view the report(s).

The first step is to create a website. I used the IIS Administrator snap-in to create new virtual folder called Performance under the default website and used the MOM Reports folder, that I created earlier, as the home directory on the MOM Reporting server. The result of this is that the subfolder for each server and the TIF files that are the results of the link reports are contained within the website.

We want the website to contain image files that are the results of the linked reports that have been executed by the shared schedules. Unfortunately, we cannot embed TIF files in a website so that they can be viewed online. They can only be downloaded, which defeats the purpose of this website. We want to be able to present the data within the page. My solution was to use a graphics conversion program to convert the TIF iages to JPEG. I can then embed the JPEGS in my website. I used a command line batch conversion program called 2Jpeg to do this. This is a shareware program but the demo program works fine. The only problem is that is puts a great big watermark all over the image.  Paying for your license gets rid of the watermark and makes you legal.  I created a batch file to run 2JPEG with the appropriate parameters and then stored the batch file in the MOM Reports folder. A scheduled task was created to run the conversion batch file at a suitable time after all of the MOM linked report subscriptions had been executed. Now, all of my TIF images were replaced by suitably sized JPEG’s.

A simple website or webpage (depending on how you want to organise things) can be created in the MOM Reports folder. The root page should be called index.htm. Edit the content of the website to put in your headers and text. Now you can link in your JPEG’s in the suitable locations. When you open the web page(s) you’ll find that your reports appear in the appropriate locations under headers and accompanied by standardised text that describe what the reports are.

How It Works

MOM will collect data as normal and export it into the MOM Reporting database using the standard DTS job that typically runs at 01:00 every morning.

Your linked reports will execute according to your shared schedules. If you have 5 reports and 100 servers then 500 reports will be run. 500 TIF images will be stored in subfolders in the MOM Reports$ share, overwriting any previously existing TIF images. The names of these files never change. I set up my example to run every Monday because I wanted to run the reports once a week.

Soon after your linked reports have finished, your scheduled task to convert the TIF files will run. You TIF files will be replaced by identically named JPEG files, e.g. “CPU Metrics.TIF” will be replaced by “CPU Metrics.JPG”.

The website page content is static. You don’t need to re-edit it every month. The web content links to the JPEG files whose names never changed but are replaced every time the linked reports are run and the batch conversion is run. A user can open this website and view the newest content as it exists in the MOM Reports$ share, i.e. the home directory of the Performance virtual folder.

Those who are responsible for generating reports can either redirect clients/management to the website. If documentation is required, a .MHT can be created (Internet Explorer – Save As) or the site can be printed out as a .PDF using a PDF printer. Either of these can be sent to clients or management.

There is no need for administrators to change anything in the configuration to run reports as normal. The only time the configuration needs to be altered is when a new managed server is added. A new subfolder in the MOM Reports$ share should be added for the server. A new subscription for each appropriate report should be added. The web content should be modified to add appropriate headers, text and to embed the new JPEG’s for the report images for this new server. That’s all there is.

Summary

So that’s my solution for producing a regular SLA report from the MOM 2005 Reporting database. It’s not perfect. I’d love some simple client where I could say “go produce a report based on this dynamic time period” and it would come back with a lovely Microsoft Word or PDF document after a period of processing. I’m sure it can be done but it would take a great deal of effort.

My solution does take a bit of effort to get going, especially in defining all of the linked report subscriptions. But, once it is up and running you’ve got a free solution that requires very little alteration and enables you to present granular and complete data in two useful formats.

Introduction to IP Version 6

Last year, I was lucky enough to be at a presentation by Todd Lammle, of Cisco literature fame, about IPv6.  We knew nothing about IPv6 going in and we came out knowing that we definitely knew nothing.  IPv6 is a whole new way of network addressing and routing.  It certainly takes a lot to get your head around it.  Microsoft have included support for IPv6 in Vista and rumour is that it was driven by a national IPv6 deployment in China.  Windows "Longhorn" will also include support.

Microsoft has released a document that gives you an introduction to this new protocol.  You should give it a read if IPv6 is i your future or even if you are just curious.  Microsoft also has a website set up to support this technology.

Credit: Bink

What’s Next?

I’ve started getting back to work on System Center Configuration Manager 2007.  The core functionality hasn’t changed much from SMS 2003 so it’s not a great leap.  Some of the 2003 feature packs are integrated but surprisingly, not all of them are. 

I’m going to start documenting the different parts of the product.  Instead of the usual one doc with 4 gazillion screenshots covering everything, I’m going to do one document for each major feature.  I’ll kick off with basic concepts, design and installation of CM 2007 (based on information available to the public).  You can expect that in the coming weeks.

Quest Software: Exchange 2007 Readiness Event

Colm Torris at TechNet Ireland has published a link to a half day Exchange 2007 event being held at the Guinness Storehouse in Dublin by Quest Software on the 21st of February.  Topics include:

  • Assessing your Exchange 2007 readiness
  • Options for migrating to Exchange 2007 (from 5.5, 2000, 2003, Notes or Groupwise)
  • Managing your new Exchange 2007 environment
  • Open lunch/questions

Quest Software are one of Microsoft’s best partners.  Some of their best known solutions are intended for migration scenarios.  You can be sure that this will focus on Exchange 2007 migration scenarios and how Quest can provide a solution for you.

Credit: TechNet Ireland.

Visual Studio “Orcas” CTP

The Community Technology Preview version of the the latest version of Visual Studio has been released.  I’ve not been a developer since my first few months out of college so it’s not of great interest to me.  But … even us infrastructure types are involved with Visual Studio in some ways.  SQL Reporting Services relies on it if you want to create custom reports and we often get involved with application deployments that involve licensing and installing it.  The latter is important because as far as I know, it’s either not possible or damned expensive to get Software Assurance for Visual Studio.  Why’s this important?  Because this "Orcas" version is what Microsoft wants you to use for coding for Vista and Office 2007.  I’ve found from personal experience that developers who stay in touch are keen to take advantage of new tools and versions of Office.  This may push the deploymenbt of Vista, Office 2007 and Vistal Studio "Orcas" for some organisations, even if on a limited basis.

Here’s an excerpt of what Microsoft has to say:

Visual Studio code name “Orcas” delivers on Microsoft’s vision of smart client applications by enabling developers to rapidly create connected applications that deliver the highest quality rich user experiences. This new version enables any size organization to rapidly create more secure, manageable, and more reliable applications that take advantage of Windows Vista and the 2007 Office System. By building these new types of applications, organizations will find it easier than ever before to capture and analyze information so that they can make effective business decisions.