Windows 2008 Failover Clustering: Configuring SQL 2008 for Failover Clustering – Part 1 of 3

Configuring Windows 2008 Failover Clustering is another mechanism we have to make Applications ‘Highly Available’ for End Users.  Many of the ‘Windows Core Infrastructure Applications’ (Operations Manager 2007, Configuration Manager 2007, etc.) require a specific version of Microsoft SQL as the Core Database Engine and/or Core Reporting Services Engine to function.  In this multi-part Blog entry I will review some of the configuration steps to offer SQL 2008 Services on a Windows 2008 Failover Cluster.

Here are the Blog entries that comprise the aggregate solution:

Windows 2008 Failover Clustering: Configuring SQL 2008 for Failover Clustering – Part 1 of 3 (this Blog entry!)

Windows 2008 Failover Clustering: Configuring SQL 2008 for Failover Clustering – Part 2 of 3

Windows 2008 Failover Clustering: Configuring SQL 2008 for Failover Clustering – Part 3 of 3

First, I will build the Windows 2008 Failover Cluster using 3 Nodes and a Node Majority for the Quorum Configuration.  Here is that completed work.


Figure A – 3 Node Windows 2008 Failover Cluster.  This is built in Blog Part 1 and 2.
Then I will build the Application Cluster running Windows 2008 Failover Clustering and SQL 2008 Failover Clustering Services.

Figure B – 3 Node Windows 2008 Failover Cluster with SQL 2008 Failover Cluster across all Nodes. This is built in Blog Part 3.



Figure 1 – I begin with 3 Nodes configured as Windows 2008 x64 Enterprise Edition Servers.  These Servers are Patched through July 2009 and include the following Feature:

Failover Clustering

In addition, I am using the StarWind iSCSI Server as the iSCSI Target and the Microsoft iSCSI Initiator application provided as part of the Operating System for Windows 2008.



Figure 2 – The ‘Disk Management’ MMC in Server Manager shows the LUNs presented over iSCSI to all three Failover Cluster Nodes.  I always configure a ‘Quorum Drive’, even for a configuration like this (with 3 Nodes) that does not require a Quorum Drive.


Figure 3 – I get underway by using Server Manager invoking the ‘Add Roles’ function.  It is my intention to add the ‘Application Server’ Role initially.



Figure 4 – The initial screen for the ‘Add Roles Wizard’ confirms details specific to proceeding.



Figure 5 – In the ‘Add Roles Wizard’ I select ‘Application Server’.  The paradigm of ‘Application Servers’ encompasses many different aspects of running an Application on the Windows 2008 Platform, and is most focused on Web Services.



Figure 6 – Upon adding the ‘Application Server’ Role the prerequisite ‘Features’ offer a Confirmation of addition of such Features.  Specifically, the ‘Application Server’ Role requires ‘Windows Process Activation’.



Figure 7 – After confirmation of the Features required by the Application Server Role the ‘Add Roles Wizard’ offers validation of the scope of Services provided by adding the ‘Application Server’ Role.



Figure 8 – One of the primary reasons for adding the ‘Application Server’ Role is addition of ‘Data Transformation Services’.  Here, I confirm addition of both ‘Incoming Remote Transactions’ and ‘Outgoing Remote Transactions’.



Figure 9 – Upon confirmation of both the Role and Features I am ready to ‘Install’ to proceed.



Figure 10 – Installation successful!  Next I will proceed to confirm the COM+ Application configuration.



Figure 11 – Here I review the COM+ Application configuration resulting from the choices for both ‘Role’ and ‘Features’ for the ‘Application Server’ Role.



Figure 12 – Previously, I installed the ‘Feature’ of ‘Failover Clustering’ that includes the ‘Failover Clustering Tools’.  This occurred while configuring the iSCI Initiator and iSCSI Target parameters.



Figure 13 – Upon opening the ‘Failover Clustering Management’ MMC, the recommended approach is to select ‘Validate a Configuration (Cluster Configuration)’ from the ‘Management’ leaf of the MMC.



Figure 14 – The ‘Validate a Configuration Wizard’ presents relevant information on proceeding through the Validation Process.



Figure 15 – In the next step of the ‘Validate a Configuration Wizard’ I input the Host by Name (b01-node-1.corp.itpslab.local) and proceed to confirmation Test Parameters.



Figure 16 – The default value of ‘Run All Tests’ is reasonable seeing this is the first Node in a multi-Node Cluster Configuration.



Figure 17 – Selecting ‘Run All Tests’ provides a Summary of both Servers (Nodes) targeted for Testing and the Test List.



Figure 18 – Upon completion of the ‘Validate a Configuration Wizard’ Summary Information is provided in both condensed and detailed format.  The detailed format is available by selecting ‘View Report’ from the Summary screen.



Figure 19 – Upon review of the ‘Validate a Configuration Wizard’ Summary Report, and in observing no Critical Errors I proceed to ‘Create a Cluster’ using the ‘Create a Cluster’ Wizard.



Figure 20 – The initial information from the ‘Create a Cluster Wizard’ introduces details pertinent to Cluster Creation.



Figure 21 – This is the first Node and as such I simply add ‘bo1-node-1.corp.itpslab.local’ to the ‘Selected Servers’ list.



Figure 22 – In a step not show here I add ‘app-cluster-01’ as the ‘Cluster Name’ (Cluster Name Object, or CNO).  Additionally, I assign the CNO an IP Address of from the IP Schema for the ‘Public Network’.  The ‘Create a Cluster Wizard’ confirms the ability to proceed to completion in creating a single Node Cluster on Windows 2008.



Figure 23 – The ‘Create Cluster’ Report provides Summary Detail on the Cluster Creation process for further review.



Figure 24 – The additional relevant Detail from the ‘Create a Cluster’ Summary Report.



Figure 25 – Moving back to the ‘Failover Cluster Management’ MMC we observe several details about this Cluster Configuration.  Node ‘b01-node-1.corp.itpslab.local’ includes a Single Node, the Quorum Configuration is ‘Node Majority’, and 2 Separate IP Networks provide Services for this Cluster Configuration.



Figure 26 – Selecting ‘Storage’ from the ‘Failover Cluster Management’ MMC indicates ‘No Storage’ selected for the Cluster Configuration.  I can easily ‘Add a Disk’ using the ‘Action Menu’ or by ‘Right Mouse Clicking’ the ‘Storage’ leaf.



Figure 27 – Selection of ‘Cluster Network 1 (the Default Name given)’ indicates this is the IP Interface for the ‘Public Network’ for this configuration.  The ‘Public Network’ provides Client Connectivity Traffic to/from the Cluster Node.



Figure 28 – The ‘Private Network’ handles the ‘Heartbeat Traffic’ between Cluster Nodes.



Figure 29 – Next, I move to Node 2 to add the Failover Cluster Services to the Cluster title ‘app-cluster-01.corp.itpslab.local’.  This second Node is configured exactly like the first Node.



Figure 30 – I begin with the ‘Validate a Configuration Wizard’ as previously completed on Node 1.



Figure 31 – Once initiated, the ‘Validate a Configuration Wizard’ provides the detail required to add the second Node to this Failover Cluster configuration.



Figure 32 – I add the current Cluster Node (bo1-node-1) and the targeted Cluster Node (b01-node-2) to this process of Validation.



Figure 33 – I select ‘Run All Tests’ and initiate the Validation process.



Figure 34 – Here is the Summary of ‘Servers to Test’ and ‘Test Selected by the User’ for the ‘Validate a Configuration Wizard’.  Output of these Tests will confirm/deny the ability to add a second Node to the Cluster.



Figure 35 – This screen capture is ‘mid process’ to show some of the many processes for Validation that occurs.  This process titled ‘Validate Multiple Arbitrators’ is focused on Disk Arbitration for the Cluster Node.



Figure 36 – Upon successful completion of the ‘Validate a Configuration Wizard’ I select ‘View Report’ to see the Validation detail.



Figure 37 – Since Cluster Node 2 (b01-node-2) was Validated correctly I decide to bring both Cluster Node 2 and Cluster Node 3 ‘into the Cluster’ simultaneously.  Of course, for Cluster Node 3 that means passing the ‘Validate a Configuration Wizard’ will be run through successful completion.  I am confident, as all three Cluster Nodes are configured exactly the same!



Figure 38 – Same process for running the ‘Validate a Configuration Wizard’ as seen in the two prior examples.



Figure 39 – I add all three Cluster Nodes (2 ‘prospective Cluster Nodes’) for Validation.



Figure 40 – Again, I select ‘Run All Tests’ for this Validation process.



Figure 41 – Upon selecting ‘Run All Tests’ I observe the ‘Servers to Tests’ and ‘Tests Selected by the User’.



Figure 42 – All three Cluster Nodes pass the Validation Tests.  I am now ready to add both Cluster Node 2 (b01-node-2) and Cluster Node 3 (b01-node-3) to the Failover Cluster Configuration.



Figure 43 – Moving back to Cluster Node 2 (b01-node-2.corp.itpslab.local) I will begin the ‘Add a Node’ process using the ‘Failover Cluster Manager’ MMC.



Figure 44 – I select ‘Manage a Cluster’, which provides the ability to ‘Add a Node’ to an existing Cluster (app-cluster-01.corp.itpslab.local).



Figure 45 – Once focused upon the Cluster (app-cluster-01.corp.itpslab.local) I use the ‘Actions’ Pane and select ‘Add Node’.



Figure 46 – The ‘Add a Node Wizard’ will allow adding more than one Node simultaneously.



Figure 47 – I add Cluster Node 2 and Cluster Node 3 (b01-node-2 and b01-node-3) simultaneously.



Figure 48 – Upon selecting ‘Next’ I am presented with the addition of the Cluster Nodes.



Figure 49 – The Cluster Nodes were added successfully with a concise Report available for review.



Figure 50 – The ‘Add Cluster Nodes’ Report summarized all Actions that comprise the Cluster Node addition.



Figure 51 – Here is the second of three pages of the ‘Add Cluster Nodes’ Report.



Figure 52 – Here is the third of three pages of the ‘Add Cluster Nodes’ Report.



Figure 53 – Using the ‘Failover Cluster Management’ MMC I focus on observing the 3 Nodes of this Cluster.  The ‘Storage’ leaf provides summarized detail of Storage specific to ‘Total Capacity’ and ‘Available Capacity’.



Figure 54 – Selecting the ‘Node’ leaf and highlighting ‘b01-node-1’ I observe the ‘Resources’ associated with that Node.  All available LUNs (Disk Drives) are ‘Online’ and waiting ‘matching’ with an Application.



Figure 55 – Upon moving to highlight ‘b01-node-2’ note that no Disk Resources are controlled by this Cluster Node.



Figure 56 – A similar observation can be made when highlighting ‘b01-node-3’ regarding Disk Resources.



Figure 57 – I select ‘Services and Applications’ where there are currently no Services or Applications defined.  We are preparing this Failover Cluster to accept SQL 2008 Failover Clustering Services.



Figure 58 – Note that selection of the Cluster Name Object (CNO) there is ‘Summary Information’ regarding the Cluster.  Here, I focus on the ‘Quorum Configuration’ which is ‘Node Majority’.  Node Majority is appropriate for a 3 Node Cluster per the defined Quorum Model requirements.



Figure 59 – Selection of the Cluster Name Object and then ‘Configure Cluster Quorum Wizard’ invokes the possible Quorum Configuration types acceptable to a Windows 2008 Failover Cluster.  The Wizard selects the ‘most suitable’ Quorum Configuration based upon Disk, Node and Network Configuration criteria.

Summary: In this Blog entry, the first of three, I follow step-by-step procedures to 1) Validate a 3 Node Cluster Configuration, and 2) Create a 3 Node Cluster Configuration using Windows 2008 Failover Clustering.  The 3 Node Cluster will be used to Install and Configure SQL 2008 Failover Clustering Services


Dear Readers,

Currently I am working as a Associate Infrastructure Manager with a multinational software house

After doing my Bachelors of Computer Science (Honors) from the Hamdard University Karachi in 2002,  Since then I have had the opportunity to keep myself up-to-date with the latest technologies and responsibilities. During the job I further enhanced my professional qualification by obtaining the MSc Computer Science degree from Government College University Lahore, which further broadened my vision and my approach to my current job responsibilities.

Best Regards


Scarred Girl in KFC Controversy

Go Fund Me Suspends Site for Scarred Girl in KFC Controversy

An online crowd funding site has suspended a campaign to raise money for a scarred girl allegedly asked to leave a Mississippi Kentucky Fried Chicken restaurant amid questions about the family’s story, and offered refunds to donors who felt they were duped.

“In lieu of the ongoing uncertainty surrounding the ‘Victoria’s Victories’ online fundraising effort, GoFundMe has temporarily suspended the campaign until the full truth is made clear,” GoFundMe CEO Brad Damphousse said in a statement Tuesday.


The story of 3-year-old Victoria Wilcher went viral after the girl’s grandmother, Kelly Mullins, claimed the child, whose face is badly scarred after a pit bull attack, was asked to leave the Jackson KFC on May 15 by an employee who said her appearance was scaring other customers.

On Tuesday, KFC said in a statement it found no evidence the incident occurred. The restaurant chain said it would still honor its pledge to donate $30,000 toward Victoria’s medical bills.

“After the alleged incident was reported to us, two investigations took place, including one by an independent investigator. Neither revealed any evidence that the incident occurred and we consider the investigation closed,” the Louisville, Kentucky, based restaurant chain said in a statement.

The GoFundMe effort raised $132,515 from 2,893 donors to pay for Victoria’s medical care before the account was suspended. The company would not say how many, if any, donors have requested a refund.

 A Facebook page set up by the family, Victoria’s Victories, has since been taken down.

The family’s attorney, Bill Kellum, said Mullins stands by her claim.

“Victoria’s family appreciates the actions of KFC in their investigation of this matter,” he said in a statement. “It is deeply disappointing that other parties have taken opportunity to attack Victoria through social and news media outlets. Victoria is an innocent child with very real physical and emotional scars.”


Reference :- (Phil Helsel)





Shots fired at Pak plane while landing in Peshawar, 1 killed

PIAPakistani security forces are searching for gunmen who attacked a passenger plane as it came in to land on Tuesday.

A woman travelling on the Pakistan International Airlines flight from Riyadh in Saudi Arabia was killed. At least two others on board were injured.

The attack has sparked renewed alarm about three weeks since gunmen stormed Karachi airport killing 29 people.

The army has responded with air strikes against the Pakistani Taliban in the tribal areas adjoining Peshawar.

The movement has threatened to mount more attacks in revenge.

Peshawar airport came under Taliban rocket attack in December 2012 – four civilians were killed. The airport is used as a base by the Pakistan Air Force as well as for commercial flights.

Flight PK756 from Riyadh in Saudi Arabia had 178 passengers on board.

It was hit by a volley of machine gun rounds as it came in to land at Peshawar. Police say one bullet narrowly missed the captain, but a Pakistani woman passenger was fatally wounded when shots penetrated the cabin.

Her nine-year-old daughter, sitting next to her, cried “my mother is dead” as security forces boarded the plane.

No-one has been detained so far and reports say flights at Bacha Khan airport have resumed.


Reports quoting police say either five or eight bullets hit the plane. Airline officials say the shots were fired from outside the airport when the jet was still some 100m (300 feet) above the ground.


The jet suffered a heavy landing, injuring a number of other passengers, but the pilot won praise for bringing the aircraft down safely and avoiding catastrophe.

PIA official Mohammad Kifayatullah Khan told Reuters he had entered the plane and saw the fatally injured woman on her seat.

He said: “All the passengers were panicked. Some of them wanted to get out as soon as possible because they were afraid of fire inside the plane.”

A deadly attack by militants on Karachi airport this month sparked an army offensive against insurgent positions in the North Waziristan region.

Pakistan’s army launched the offensive more than a week ago, bombarding Taliban targets in the tribal area.

Some 450,000 people have since been displaced, many arriving at the town of Bannu in overloaded vehicles.

Local officials say they are doing everything they can to deal with an unfolding humanitarian crisis.

The offensive began after the attack on Karachi’s Jinnah international airport, which an Uzbek militant group and the Pakistani Taliban said they had carried out.

The attack signalled an end to an already faltering peace process between the government and the Pakistani Taliban.

In another sign of the militants fighting back, Tuesday saw Pakistan’s suicide bombing since the offensive began. The bomber killed two soldiers and a civilian near a checkpoint at Spinwam in North Waziristan.



Debug Diagnostic 2.0 – Creating a rule in crash mode

This article will detail the action plan you can implement to create a rule in Debug Diagnostic 2.0 to automate a dump generation when an IIS process is crashing (in example a “W3WP.exe” process).

This action plan can be applied when your W3WP.exe process is crashing which means, in an IIS context, the specified process has been killed or restarted for an obscure reason. This implies a stop of the process and if it has been restarted, a change in the Process ID. You really have to make the difference between a crash and a hang, as configuring a crash rule if you are facing a hang will not generate any dumps…

Action Plan

If the operating system is 32-bit, install the 32 bit DebugDiagx86.msi. For 64 bit operating system choose the DebugDiagx64.msi.

  • Execute DebugDiag 2.0 Collection on the IIS server, the wizard “Select Rule Type” loads
  • Select “Crash” and then click on “Next >”


  • In the window “Select Target Type”
    • Check “A specific IIS web application pool”, then click on “Next >”
      This will have for effect to only generate a dump on the W3WP.exe process corresponding to your Application Pool

      • If you want to monitor every processes owned by IIS, you can select “All IIS/COM+ related processes”
      • You can also select “A specific process” to monitor every instances of a process or just a unique one (a specific process with a specific PID)
        The drawback of “A specific process” for a unique instance is you need to recreate the rule each time the process is restarted as his PID would have changed.




  • Select the Application Pool facing the crash issue, then click on “Next >”



  • Add the breakpoint as you can see in the below screenshot
    • This allows you to generate a dump as soon as the command “TerminateProcess” is sent in the process without waiting for a complete stop
    • In addition, if a problem is detected by WAS in the process, it will stop it with this command, so without this breakpoint, no dump will be generated while the process is facing a crash



  • Then click on “Next >”



  • Choose a name and a location for dumps file, then click on “Next >”
    • Setting those dumps on another disk than C:\ is possible
    • Please check you have enough available space disk as the dump size will be equal to the memory consumed by the process you are monitoring




  • In the window “Rule Completed”, select “Activate the rule now”, and then click on “Finish”
    • The rule is now configured and ready to generate dumps




  • Check the status is “Active”
    • If it’s not, The “DbgSvc” service is surely not started
    • You can go in “Administrative Tools > Services” and start it
  • When a dump will be generated, you’ll see the column “UserDump Count” being incremented

The dumps are located by default in “C:\Program Files\DebugDiag\Logs\Crash rule for all IIS_COM+ related processes” or in “C:\Program Files\DebugDiag\Logs\Rule_Name”.


If you see the following dialog box, choose No because we only need to generate dump files.



Author :- Xin Jin From GBSD DSI Team