Welcome to TechNet Blogs Sign in | Join | Help

The presentations from SNIA's SDC 2009 are now publicly available for download. This SNIA event happened in Santa Clara, CA last September and the decks are now publicly available, after being restricted to conference attendees for a few months.

Out of the dozens of presentations, I wanted to highlight a few that were delivered by people from Microsoft:

ATA Trim/Delete Notification Support in Windows 7
Neal Christiansen
http://www.snia.org/events/storage-developer2009/presentations/thursday/NealChristiansen_ATA_TrimDeleteNotification_Windows7.pdf

BranchCache: Next Generation Branch Office Optimizations
Ravi Rao, Molly Brown
http://www.snia.org/events/storage-developer2009/presentations/monday/RaviRao-MollyBrown_BranchCache.pdf

Caching SMB Data for Offline Access and an Improved Online Experience
Molly Brown
http://www.snia.org/events/storage-developer2009/presentations/monday/RaviRao-MollyBrown_BranchCache.pdf

DFS-N Overview and Scalability
Marcello Hasegawa, Saad Ansari, Jose Barreto
http://www.snia.org/events/storage-developer2009/presentations/wednesday/SaadAnsari-Hasegawa_Barreto_DFS-N_Overview-rev.pdf

File Server Capacity Tool (FSCT) for CIFS/SMB/SMB2
Jose Barreto, Bartosz Nyczkowski, Jian Yan
http://www.snia.org/events/storage-developer2009/presentations/tuesday/BartoszNyczkowski-JianYan_FileServerCapacityTool.pdf

Green Storage in the Data Center
SW Worth
http://www.snia.org/events/storage-developer2009/presentations/monday/SWorth-GreenStorage_SDC_2009_2.pdf

Managing Storage of Virtual Machines with Microsoft System Center Virtual Machine Manager (SC-VMM)
Madhu Jujare
http://www.snia.org/events/storage-developer2009/presentations/tuesday/MadhuJujare_ManagingStorage_Virtual_Machines.pdf

Microsoft Network Monitor 3 (NM3)
Paul Long
http://www.snia.org/events/storage-developer2009/presentations/wednesday/PaulLong_TShootSMBwithNM3-rev.pdf

Microsoft Storage Management Update
Mohamed Lawindi, Michael Brasher
http://www.snia.org/events/storage-developer2009/presentations/wednesday/Mlawindi_MikeBrasher_Microsoft_StorMgmt_v1.pdf

SMB v2.1
David Kruse
http://www.snia.org/events/storage-developer2009/presentations/tuesday/DavidKruse_SMBv21.pdf

SMB2 Model Based Testing
Nico Kicillof, Albert Lee
http://www.snia.org/events/storage-developer2009/presentations/tuesday/NicoKicillof-AlbertLee_SMB2ModelBasedYesting.pdf

Understanding Windows File System Transactions
Christian Allred
http://www.snia.org/events/storage-developer2009/presentations/tuesday/ChristianAllred_UnderstandingWindowsFileSystemTransactions.pdf

Using Data Classification to Manage File Servers
Adi Oltean
http://www.snia.org/events/storage-developer2009/presentations/tuesday/AdiOltean_UsingDataClassificationToManageFileServers.pdf

I would also encourage you to check the full list of presentations from SDC 2009, visiting the site at
http://www.snia.org/events/storage-developer2009/presentations/

I get a number of questions regarding reference information for DFS Namespaces (also known as DFS-N). I gathered here the main links that can get you started on the subject:

Here are my own blog posts on DFS Namespaces (all updated for Windows Server 2008 R2 and including domain V2 namespaces):

Next, here are a couple of public DFS Namespaces Presentations by Microsoft:

Also, here are the main TechNet Reference links on DFS Namespaces for Windows Server 2008 and Windows Server 2008 R2 (includes domain V2 namespaces):

Last, here are some TechNet Reference on DFS Namespaces for Windows Server 2003 (does not include domain V2 namespaces): 

I hope those help you with your DFS Namespaces projects.

The Windows Storage Server Team just released a new 30-page white paper on "Configuring Failover Clusters with Windows Storage Server 2008".

This paper will guide you through the process of configuring the networks, domains and clustering features, including performance recommendations for a pair of Windows Storage Server 2008 appliances hosting a File Server or a Microsoft iSCSI Software Target in a high availability failover cluster.

The table of contents for this new white paper shows:

  • Failover Cluster Prerequisites
    • Establish a Network Naming Convention   
    • TCP/IP Network Configuration   
    • Public Network
    • Storage Network
    • Heartbeat Network   
  • Procedures
    • Prepare the Failover Cluster
    • Create a Domain User Account
    • Add Nodes to an Active Directory Domain
    • Expose Storage to Cluster Nodes
    • Install the Failover Cluster Feature
    • Run Cluster Validation
    • Create and Configure the Failover Cluster
    • Create a Cluster
    • Set Cluster Network Properties and Apply Naming Convention
  • Create a Highly Available File Server
    • Mapping User Folders to the Highly Available File Server Share
  • Create a Highly Available iSCSI Target
    • Configuring Windows Firewall for Microsoft iSCSI Software Target
    • Installing the Microsoft iSCSI Software Target
    • Create the Failover iSCSI Target Resource Group
    • Create an iSCSI Target in the Microsoft iSCSI Target MMC
    • Create and Configure Virtual Disks
    • Connect Initiators
  • Microsoft iSCSI Software Target Performance Recommendations
  • Testing Your Failover Cluster Configuration

Download it from
http://download.microsoft.com/download/3/B/5/3B5632C8-7A04-44A6-8ED2-A122C2D6DDB1/ConfigureFailover.docx

Windows Server 2008 R2 introduced a new File Classification Infrastructure (FCI) and the team has been sharing different ways to put it to good use.

The latest one comes as a blog post that describes how to use the FCI SharePoint Upload PowerShell script in conjunction with the FCI File Management Tasks to automatically upload files from a File Server (running Windows Server 2008 R2) to SharePoint (both SharePoint Server 2007 and SharePoint Server 2010 Beta) based on a specific condition.

The post outlines the motivations and the process to automate the move of certain files from a file share to a document library. It also covers some more complex decisions you'll have to make in the process, like whether to delete the file from the file share or how the classification atributes can be preserved after the upload to SharePoint. This is guaranteed to get you thinking about a whole new set of scenarios that involve both File Servers and SharePoint Servers.

Read the full post at http://blogs.technet.com/filecab/archive/2009/12/14/automatically-upload-files-from-file-server-to-sharepoint-using-the-file-classification-infrastructure-fci.aspx

Scott M. Johnson, a Program Manager in the Windows Storage Server team, has just published a new blog post that might give you some ideas on different ways to use the Microsoft iSCSI Software Target. He outlines six different uses for it:

1) Consolidate storage for multiple application servers.
2) Test and Development scenarios are endless, especially for Clustering, Live Migration, SAN transfer and Storage Manager for SANs.
3) Setup an iSCSI SAN for a Windows cluster.
4) Consolidate servers into Hyper-V VMs and migrate the data to a Windows Storage Server. 
5) Diskless SAN boot over iSCSI!
6)  “Bonus storage” for people in your organization. Storage Administrators can be a hero! (for once).

See the full post (with lots of details and diagrams) at:
http://blogs.technet.com/storageserver/archive/2009/12/11/six-uses-for-the-microsoft-iscsi-software-target.aspx

A while back, Microsoft announced the release of PowerShell v2 for Windows XP, Windows Server 2003, Windows Vista, and Windows Server 2008 (see http://go.microsoft.com/fwlink/?LinkID=151321).

However, it is not clear to everyone that Powershell v2 is already part of Windows 7 and Windows Server 2008 R2.

I have seen people still asking for the link for PowerShell v2 for Windows 7 and Windows Server 2008 R2. Such download does not exist, since the PowerShell included in the OS is already v2.

Part of this confusion may be due to the fact that the folder in Windows 7 is still called "C:\windows\System32\WindowsPowerShell\v1.0". Maybe it's the fact that we still use the "PS1" file extension. Maybe it's the fact that the file version number for PowerShell.exe is actually "6.1.7600".

Just believe that v2 is already there... OK, maybe you don't believe me. Here's a little cmdlet to confirm it for you. You can launch PowerShell and use "get-host".

That should make it perfectly clear :-)

Great news following a very busy week for SQL Server with the PASS Community Summit: The new SQL Server 2008 R2 November CTP is now available for MSDN and TechNet subscribers.This download is expected to be available as a general download in a couple of days.

Title: SQL Server 2008 R2 Enterprise Evaluation November CTP (x86, x64, ia64) - DVD (English)
Size: 4,011.88 (MB)
File Name: en_sql_server_2008_r2_enterprise_evaluation_november_ctp_x86_x64_ia64_dvd_455348.iso 
Date Posted: 11/9/2009

SQL Server 2008 R2 Nov CTP

I'm already downloading... :-)

Additional details at
http://www.microsoft.com/presspass/press/2009/nov09/11-03pass09pr.mspx

I came across some interesting tips on things to consider in a virtualized environment by Greg Shields (Consultant and Microsoft MVP). 
They are non-issues (or settled issues) on physical environments and you might overlook them when making the transition to virtual machines.

Here's a summary to get you interested:

  • Screensavers
  • Managing from the console
  • AV on VM disk files
  • Power options
  • Cluster Failback
  • RAM availability
  • Backup for CSV
  • Number of processors

Here are the links to the actual articles:

This post mixes the spirit of Halloween and the passion for SQL Server :-)

 

Have you noticed how SQL Server has a number of things that seem right out of a horror movie?

As a test of your SQL Server expertise and for a fun Halloween trivia game, try to describe the items in the list below in the context of SQL Server.

If you can’t, follow the links to figure them out:

And, of course, there’s a number of Wizards all over the place.

If you know any other good ones, please do share in the post comments.

 

P.S.: I was also reminded by Kalen Delaney of the classic "Halloween Problem" which can cause pretty scary results, as described at http://en.wikipedia.org/wiki/Halloween_Problem. Not specific to SQL Server, but there was a related issue with SQL Server 7 : http://support.microsoft.com/kb/248441.

 

Happy Halloween!

SQL Halloween

There's a new 77-page white paper covering the use of Folder Redirection and Offline Files in a very practical way.

Here's an excerpt:

"The objective of this white paper is to show through a case study how to use different Microsoft products and technologies to put in place a comprehensive solution satisfying the needs of a mid-sized organization around users’ file data management. The study was conducted by the Quality Assurance group of the Storage Solutions Division (SSD) at Microsoft, a division that focuses on enabling customers of all sizes to store, manage, and reliably access their file data."

Download from http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=d8541618-5c63-4c4d-a0fd-d942cd3d2ec6

Jeff Teper (Corporate Vice President, SharePoint Server) and Steve Ballmer (CEO) shared SharePoint Server 2010 feature details and announced the November beta during the SharePoint Conference today in Las Vegas.

Jeff also posted a long and revealing blog, outlining 8 categories and 40 feature areas:

  • Sites
    • SharePoint Web Experience
    • Office Client
    • SharePoint Workspace
    • Office Web Apps
    • SharePoint Mobile Access
  • Communities
    • Collaborative Content
    • Social Feedback and Organization
    • User Profiles
    • MySites
    • People Connections
  • Content
    • Large Lists and Libraries
    • Enterprise Metadata
    • Document Sets
    • Web Publishing including Digital Asset Management
    • Governance and Records Management
  • Search
    • Interactive Search Experience
    • Relevance
    • People Search
    • Connectivity
    • Scale and Platform Flexibility 
  • Insights
    • Excel Services
    • Performance Point Services
    • SQL Server
    • “Gemini”
    • Visio Services
  • Composites
    • SharePoint Designer
    • InfoPath Forms Service
    • Access Services
    • Sandbox Solutions
    • Business Connectivity Services
  • Administration
    • Improved Upgrade
    • Throttling, Health Monitoring, Analytics
    • Web and PowerShell Admin
    • Scalability and Availability
    • Identity Management and Security
  • Development
    • New SharePoint APIs
    • Application Lifecycle
    • Visual Studio 2010 Support
    • Developer Dashboard View
    • Development on Windows 7

Jeff did an amazing job summarizing the upcoming SharePoint Server 2010 in a single blog post that includes additional information on each of the 48 bullets above, plus 8 screenshots (one per category). Absolute must read!!!  

Check the blog post at

Recorded keynote from the SharePoint Conference available from

Also, documentation and learning resources on SharePoint 2010 are available starting today for both Developers and IT Professionals:

SharePoint 2010

Version 1.0 of the File Server Capacity Tool (FSCT) was announced yesterday during a presentation by Jian Yan and  Bartosz Nyczkowski at SNIA’s Storage Developer Conference in Santa Clara, CA. The presentation covered a number of details about FSCT and included a demo running FSCT with the HomeFolders workload.

If you are not familiar with FSCT, the download page offers this overview: “File server capacity planning and performance troubleshooting are critical aspects of high-level network administration. Central file servers and distributed client workstations are now the norm in most corporate networks. This structure reduces storage capacity requirements, centralizes backup, increases the availability of files, and simplifies the document revision and review process. However, because data storage and access are centralized, performance limitations impact the entire network population. Accurately projecting the number of users that hardware can support under a specific workload, and understanding when and where bottlenecks occur, are critical to making efficient improvements to the server configuration. File server capacity planning tools can be valuable in choosing new hardware for purchase, identifying the capacity of existing hardware, locating existing bottlenecks, and planning for resource expansion in advance of resource exhaustion. The throughput capacity of a file server can be expressed either as the maximum number of operations per second or a maximum number of users supported by the configuration. These values are influenced by several factors, some of which include processor speed, available memory, disk speed, network throughput and latency, and the speed with which SMB requests are processed.”

The final version is available for download in both 32-bit (x86) and 64-bit (x64) versions. It is supported on  Windows Server 2003, Windows Server 2003 R2, Windows Server 2008, Windows Server 2008 R2, Windows Vista and  Windows 7 (with the latest service pack applied).

Downloads are available now:
x64: http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=b20db7f1-15fd-40ae-9f3a-514968c65643
x86: http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=0b212272-1884-4af1-972d-42ef1db9f977

A white paper is also included in the download packages, with detailed description of the tool, step-by-step instructions on how to use it and reference of the command line interface.

For questions about FSCT and how to use it, please use the forum at
http://social.technet.microsoft.com/Forums/en-US/fsct/threads

Also check some details on FSCT on these previous posts (from the release of the beta last year and the release candidate back in July):
http://blogs.technet.com/josebda/archive/2009/07/08/file-server-capacity-tool-fsct-release-candidate-available-for-download.aspx
http://blogs.technet.com/josebda/archive/2008/10/06/fsct-a-cifs-smb-smb2-file-server-tool-for-capacity-planning-and-performance-troubleshooting.aspx

This is a major milestone in the two-year journey to offer this tool publicly, which included efforts from a number of people from different teams at Microsoft, including the File Server Team and the Windows Performance Team.

I recently bought a used TI-55-II on eBay. This is an old (early 80’s) Texas Instruments calculator, one the first programmable ones. I got it mostly for sentimental value, since I used to own one back in Brazil in 1983, my last year in high school.

Looking back, the fact that I got this specific programmable calculator exactly at that moment in my life was probably an important contributing factor to my choice of Computer Science when I joined the Federal University of Ceara in Brazil in 1984. Before that, I had my eyes set on Architecture. And not Computer or Systems Architecture, I should say, since back then we did not associate Architecture with Computer Science careers.

The TI-55-II is actually a pretty limited device by today’s standards. You have a 10-digit LCD display, 8 “memories” and 56 “programming steps”. Programming the calculator basically meant storing a sequence of keystrokes and you had no conditional statements, just one RST command to go back to step 0. Even with that, I remember being quite impressed with the ability to create a program, use multiple variables and display data on the screen (there is a PAUSE instruction to let you see a number on the screen before moving to the next step).

The calculator I used back in high school was actually manufactured in the city of Manaus in Brazil ("Produzida na Zona Franca de Manaus") by a subsidiary of Texas Instruments. I found some information about it on the internet: http://www.datamath.org/Sci/Slanted/TI-55-II-AA.htm. In fact, there is even a soft copy of the manual in PDF format: http://www.datamath.net/Manuals/TI-55-II_QR_US.pdf.

My used TI-55-II I got from eBay (shown on the right) was made in the US in 1982. It's almost the same as the Brazilian model except for the type of battery it uses. Details at http://www.datamath.org/Sci/Slanted/TI-55-II.htm. Mine is actually working fine and I even managed to remember how to create simple programs with it. Interesting how your brain can retain that kind of information decades later…

See also a previous blog post on CP/M and the TRS-80, which I used a few years after that: http://blogs.technet.com/josebda/archive/2009/08/07/the-good-old-days-of-cp-m-2-2-on-a-trs-80-with-an-8-bit-z80-cpu.aspx

 TI-55-II

The DFS-N test team has completed some extensive Performance and Scalability testing and we wanted to share some of the results.

It includes results on standalone namespaces, "2000 mode" domain namespaces and "2008 mode" domain namespaces, under Windows Server 2003, Windows Server 2008 and Windows Server 2008 R2. 

Check the post by Marcello Hasegawa at http://blogs.technet.com/filecab/archive/2009/08/22/windows-server-dfs-namespaces-performance-and-scalability.aspx.

Introduction

In this blog post, we’ll showcase the flexibility you have with DFS-N by showing three distinct ways that you could design a namespace.

Note: If you’re not familiar with DFS-N, please read this blog post first:
http://blogs.technet.com/josebda/archive/2009/03/10/the-basics-of-the-windows-server-2008-distributed-file-system-dfs.aspx

Scenario

The scenario here is simple: you have three file servers in different locations: S1 is in New York, S2 is in Amsterdam and S3 is in Hong Kong. Each server has a unique set of data (Proposals, Marketing information and Engineering documents for each geography) and you want to expose this under a single namespace. Here’s the structure on each file server:

Server S1 in New York:

  • F:\AMER\PROPOSAL – Proposals for the AMER (Americas) region
  • F:\AMER\MARKETING – Marketing information for the AMER region
  • F:\AMER\ENGINEERING – Engineering documents for the AMER region

Server S2 in Amsterdam:

  • F:\EMEA\PROPOSAL – Proposals for the EMEA (Europe, Middle East and Africa) region
  • F:\EMEA\MARKETING – Marketing information for the EMEA region
  • F:\EMEA\ENGINEERING – Engineering documents for the EMEA region

Server S3 in Hong Kong:

  • F:\APAC\PROPOSAL – Proposals for the APAC (Asia and Pacific) region
  • F:\APAC\MARKETING – Marketing information for the APAC region
  • F:\APAC\ENGINEERING – Engineering documents for the APAC region

Please note that this specific scenario does not include replication. Each server each has completely an independent set of files. For instance, a proposal for an EMEA customer would exist only in the F:\EMEA\PROPOSALS folder in S2 and it would not exist in servers S1 or S3. Replication across these servers could be accomplished using DFS-R, but I am purposely avoiding it in this example to focus specifically on DFS-N design.

Creating the folders and shares

First of all, I am providing the scripts to create the folder structure and the shares. These are 3 distinct scripts and each one needs to be run on a specific server (S1, S2 or S3). We could actually do this all remotely, but I’m keeping it simple.

CREATE-S1.CMD

REM ## RUN THIS SCRIPT in S1, the file server in New York
MD F:\AMER\PROPOSALS
MD F:\AMER\MARKETING
MD F:\AMER\ENGINEERING
NET SHARE AMER=F:\AMER

CREATE-S2.CMD

REM ## RUN THIS SCRIPT in S2, the file server in Amsterdam
MD F:\EMEA\PROPOSALS
MD F:\EMEA\MARKETING
MD F:\EMEA\ENGINEERING
NET SHARE EMEA=F:\EMEA

CREATE-S3.CMD

REM ## RUN THIS SCRIPT in S3, the file server in Hong Kong
MD F:\APAC\PROPOSALS
MD F:\APAC\MARKETING
MD F:\APAC\ENGINEERING
NET SHARE APAC=F:\APAC

Please note that the shares will end up with read-only permissions using the commands above. In your real-world deployment you will need to properly plan your NTFS and file share permissions and implement them properly with CACLS or ICACLS.EXE (for NTFS) and the /GRANT option in NET SHARE (for the file shares). The focus of this post is not on permissions but on the structure of the shares, folders and namespaces.

Namespace option 1 – Simple mapping of shares

The first option I will cover is a simple mapping at the root share on each server. This is simpler because it requires a namespace with only three folders (or links). Assuming we use a fourth server (S4) as the namespace server (or target) and NS1 as the name of the namespace (or root), users would basically use \\S4\NS1 as the UNC path to the entire set of data from all three file servers.

Note 1: You could host the namespace in one of the file servers, see details at  http://blogs.technet.com/josebda/archive/2009/06/26/how-many-dfs-n-namespaces-servers-do-you-need.aspx
Note 2: We’re using a standalone namespace here, but you could just as easily use a domain namespace instead.
Note 3: We’re not adding any fault tolerance for the namespace or the file servers in these scenarios. This could be accomplished with Failover Clustering for standalone namespaces or with multiple targets for a domain namespace.

Here’s a script to create the namespace, which should be run on S4:

CREATE-NS1.CMD

REM ## scenario 1 - Simple mapping of shares
REM ## RUN THIS SCRIPT in S4, the namespace server
MD F:\NS1
NET SHARE NS1=F:\NS1
DFSUTIL ROOT ADDSTD \\S4\NS1
DFSUTIL LINK ADD \\S4\NS1\AMER \\S1\AMER
DFSUTIL LINK ADD \\S4\NS1\EMEA \\S2\EMEA
DFSUTIL LINK ADD \\S4\NS1\APAC \\S3\APAC
DIR \\S4\NS1 /S /B
DFSUTIL ROOT EXPORT \\S4\NS1 F:\NS1EXPORT.XML VERBOSE

From an end-user perspective, the folder structure under \\S4\NS1 would look like this:

DFS NSD 1

At the end of this script, an XML export of the namespace is created. Here’s what it looks like:

NS1EXPORT.XML

<?xml version="1.0"?>
<Root xmlns = "http://schemas.microsoft.com/dfs/2007/03/dfsutil" majorVersion = "2" minorVersion = "0"
    Name="\\S4\NS1" State="OK" Timeout="300" >
    <Target State="ONLINE" >\\S4\NS1</Target>
    <Link Name="AMER" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S1\AMER</Target>
    </Link>
    <Link Name="EMEA" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S2\EMEA</Target>
    </Link>
    <Link Name="APAC" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S3\APAC</Target>
    </Link>
</Root>

Namespace option 2 – Flattening out the tree in the namespace

This second options shows the flexibility of DFS-N by creating a flat list of folders from the three servers and their folder structures. Users might find convenient to see a long list of folders without any nesting of folders. This shows how you can point your folder targets (or link targets) to a folder inside the share, not only the root of the file share. Again we’ll use that fourth server (S4) as the namespace server (or target), but now with NS2 as the name of the namespace (or root). Users would use \\S4\NS2 as the UNC path to the entire set of data from all three file servers.

Here’s a script to create the namespace, which should be run on S4:

CREATE-NS2.CMD

REM ## scenario 2 - Flattening out the tree in the namespace
REM ## RUN THIS SCRIPT in S4, the namespace server
MD F:\NS2
NET SHARE NS2=F:\NS2
DFSUTIL ROOT ADDSTD \\S4\NS2
DFSUTIL LINK ADD \\S4\NS2\AMERPROP \\S1\AMER\PROPOSALS
DFSUTIL LINK ADD \\S4\NS2\AMERMARK \\S1\AMER\MARKETING
DFSUTIL LINK ADD \\S4\NS2\AMERENGI \\S1\AMER\ENGINEERING
DFSUTIL LINK ADD \\S4\NS2\EMEAPROP \\S2\EMEA\PROPOSALS
DFSUTIL LINK ADD \\S4\NS2\EMEAMARK \\S2\EMEA\MARKETING
DFSUTIL LINK ADD \\S4\NS2\EMEAENGI \\S2\EMEA\ENGINEERING
DFSUTIL LINK ADD \\S4\NS2\APACPROP \\S3\APAC\PROPOSALS
DFSUTIL LINK ADD \\S4\NS2\APACMARK \\S3\APAC\MARKETING
DFSUTIL LINK ADD \\S4\NS2\APACENGI \\S3\APAC\ENGINEERING
DIR \\S4\NS2 /S /B
DFSUTIL ROOT EXPORT \\S4\NS2 F:\NS2EXPORT.XML VERBOSE

From an end-user perspective, the folder structure under \\S4\NS2 would be flat, like this:

DFS NSD 2

At the end of this script, an XML export of the namespace is created. Here’s what it looks like:

NS2EXPORT.XML

<?xml version="1.0"?>
<Root xmlns = "http://schemas.microsoft.com/dfs/2007/03/dfsutil" majorVersion = "2" minorVersion = "0"
    Name="\\S4\NS2" State="OK" Timeout="300" >
    <Target State="ONLINE" >\\S4\NS2</Target>
    <Link Name="EMEAMARK" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S2\EMEA\MARKETING</Target>
    </Link>
    <Link Name="APACENGI" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S3\APAC\ENGINEERING</Target>
    </Link>
    <Link Name="APACMARK" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S3\APAC\MARKETING</Target>
    </Link>
    <Link Name="AMERMARK" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S1\AMER\MARKETING</Target>
    </Link>
    <Link Name="AMERENGI" State="OK" Timeout="300" >
       <Target State="ONLINE" >\\S1\AMER\ENGINEERING</Target>
    </Link>
    <Link Name="AMERPROP" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S1\AMER\PROPOSALS</Target>
    </Link>
    <Link Name="APACPROP" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S3\APAC\PROPOSALS</Target>
    </Link>
    <Link Name="EMEAPROP" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S2\EMEA\PROPOSALS</Target>
    </Link>
    <Link Name="EMEAENGI" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S2\EMEA\ENGINEERING</Target>
    </Link>
</Root>

Namespace option 3 -  Namespace reversing the original tree structure on each file server

Last but not least, the most interesting exhibit of the flexibility of DFS-N. In this case we’ll create a namespace showing the collection of folders group by type instead of geography. DFS-N makes it easy, since we can basically restructure the shares as we create the namespace. This shows how you can also use a tree structure on the DFS folder (or link) side of the equation, which can do a lot when you combine with what we did in option 2 above. Again we’ll use that fourth server (S4) as the namespace server (or target), but now with NS3 as the name of the namespace (or root). Users would use \\S4\NS3 as the UNC path to the entire set of data from all three file servers.

Here’s a script to create the namespace, which should be run on S4:

CREATE-NS3.CMD

REM ## scenario 3 –  Namespace reversing the original tree structure on each file server
REM ## RUN THIS SCRIPT in S4, the namespace server
MD F:\NS3
NET SHARE NS3=F:\NS3
DFSUTIL ROOT ADDSTD \\S4\NS3
DFSUTIL LINK ADD \\S4\NS3\PROPOSALS\AMER \\S1\AMER\PROPOSALS
DFSUTIL LINK ADD \\S4\NS3\MARKETING\AMER \\S1\AMER\MARKETING
DFSUTIL LINK ADD \\S4\NS3\ENGINEERING\AMER \\S1\AMER\ENGINEERING
DFSUTIL LINK ADD \\S4\NS3\PROPOSALS\EMEA \\S2\EMEA\PROPOSALS
DFSUTIL LINK ADD \\S4\NS3\MARKETING\EMEA \\S2\EMEA\MARKETING
DFSUTIL LINK ADD \\S4\NS3\ENGINEERING\EMEA \\S2\EMEA\ENGINEERING
DFSUTIL LINK ADD \\S4\NS3\PROPOSALS\APAC \\S3\APAC\PROPOSALS
DFSUTIL LINK ADD \\S4\NS3\MARKETING\APAC \\S3\APAC\MARKETING
DFSUTIL LINK ADD \\S4\NS3\ENGINEERING\APAC \\S3\APAC\ENGINEERING
DIR \\S4\NS3 /S /B
DFSUTIL ROOT EXPORT \\S4\NS3 F:\NS3EXPORT.XML VERBOSE

From an end-user perspective, the folder structure under \\S4\NS3 is shown below. It examplifies how DFS-N can completely mask the physical infrastructure behind the namespace:

DFS NSD 3

Note that \\S4\NS3\ENGINEERING, \\S4\NS3\MARKETING and \\S4\NS3\PROPOSALS don’t actually exist as DFS folders (or links) at all. They are basically abstract constructs shown only in the UI, which don’t map to any specific target. You can confirm this when you look at the export of the namespace:

NS3EXPORT.XML

<?xml version="1.0"?>
<Root xmlns = "http://schemas.microsoft.com/dfs/2007/03/dfsutil" majorVersion = "2" minorVersion = "0"
    Name="\\S4\NS3" State="OK" Timeout="300" >
    <Target State="ONLINE" >\\S4\NS3</Target>
    <Link Name="MARKETING\APAC" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S3\APAC\MARKETING</Target>
    </Link>
    <Link Name="MARKETING\AMER" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S1\AMER\MARKETING</Target>
    </Link>
    <Link Name="ENGINEERING\EMEA" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S2\EMEA\ENGINEERING</Target>
    </Link>
    <Link Name="MARKETING\EMEA" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S2\EMEA\MARKETING</Target>
    </Link>
    <Link Name="ENGINEERING\AMER" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S1\AMER\ENGINEERING</Target>
    </Link>
    <Link Name="PROPOSALS\AMER" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S1\AMER\PROPOSALS</Target>
    </Link>
    <Link Name="PROPOSALS\APAC" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S3\APAC\PROPOSALS</Target>
    </Link>
    <Link Name="PROPOSALS\EMEA" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S2\EMEA\PROPOSALS</Target>
    </Link>
    <Link Name="ENGINEERING\APAC" State="OK" Timeout="300" >
        <Target State="ONLINE" >\\S3\APAC\ENGINEERING</Target>
    </Link>
</Root>

Conclusion

I hope this helped you understand how DFS-N folder (or links) and folder targets (link targets) work. Please note that we could create this all using the DFS Management graphical user interface (or MMC). We did this in the command line because it’s easier to document that way. Here’s a view from the MMC of all the three namespaces we created:

DFS NSD 4

Please note that you do not want to create all three namespaces simultaneously. I did this just for showing it here. You would choose the option that best fits your purpose. The goal here is to simplify things for the users, not to confuse them :-)

I also did not cover replication here at all. In fact, this example uses only DFS-N and can even be run without the DFS-R role service installed. We could certainly design a namespace that takes replication into account, but that’s a topic for another blog post…

Links

For more information about DFS-Namespaces, see also the following links from TechNet:

More Posts Next page »
 
Page view tracker