RyanW's questions

When provisioning with ansible in Vagrant, I am unable to get server/role specific playbooks to run. There is a common playbook run for all hosts, and that is working, but when trying to target a specific role, I'm missing something. Any ideas what piece is missing?

The ansible output is:

Running Ansible playbook...

PLAY [all] ********************************************************************


GATHERING FACTS ***************************************************************

ok: [127.0.0.1]

TASK: [common | ensure apt-get is up to date] *********************************

ok: [127.0.0.1]

TASK: [common | ensure build-essential installed] *****************************

changed: [127.0.0.1]

TASK: [common | ensure git is installed] **************************************

changed: [127.0.0.1]

TASK: [common | ensure curl is installed] *************************************

changed: [127.0.0.1]

PLAY [zookeeper_servers] ******************************************************

skipping: no hosts matched

PLAY [kafka_servers] **********************************************************

skipping: no hosts matched

PLAY [storm_servers] **********************************************************

skipping: no hosts matched

PLAY RECAP ********************************************************************

127.0.0.1                  : ok=5    changed=3    unreachable=0    failed=0

With the following directory structure:

- Vagrantfile
  - vagrant
    - playbook.yml
    - hosts
    - roles
      - common
        - tasks
          - main.yml
      - kafka_servers
        - tasks
          - main.yml
      - storm_servers
        - tasks
          - main.yml
      - zookeeper_servers
        - tasks
          - main.yml

# The main playbook

- hosts: all
  roles:
   - common

- hosts: zookeeper_servers
  roles:
  - { role: zookeeper_servers }

- hosts: kafka_servers
  roles:
  - { role: kafka_servers }

- hosts: storm_servers
  roles:
  - { role: storm_servers }

The hosts file contains

[zookeeper_servers]
zookeeper1 zoo_id=1

[storm_servers]
storm1

[kafka_servers]
kafka1

And the Vagrant file contains

# -*- mode: ruby -*-
# vi: set ft=ruby :

CUSTOM_CONFIG = {
                  "BOX_NAME"  =>  "precise64",
                  "BOX_URL"   =>  "http://files.vagrantup.com/precise64.box",
                  "HEADLESS"  =>  false
                }

Vagrant.configure("2") do |config|
  config.vm.define "kafka" do |kafka|
    kafka.vm.box = CUSTOM_CONFIG['BOX_NAME']
  end

  config.vm.define "zookeeper" do |zookeeper|
    zookeeper.vm.box = CUSTOM_CONFIG['BOX_NAME']
  end

  config.vm.define "storm" do |storm|
    storm.vm.box = CUSTOM_CONFIG['BOX_NAME']
  end

  if (/cygwin|mswin|mingw|bccwin|wince|emx/ =~ RUBY_PLATFORM)
    # provisioning with ansible on windows
    config.vm.provision "shell", path: "./vagrant/ansible-windows.sh"
  else
    # provisioning with ansible
    config.vm.provision :ansible do |ansible|
      ansible.playbook = "./vagrant/playbook.yml"
      ansible.inventory_path = "./vagrant/roles"
    end
  end
end

There are many services that advertise themselves as providing multifactor authentication for Citrix. But, most seem to require the additional generated passcode every time. Are there any that can require the password once per device, like Google's 2 step verification? So that once a device is registered, it only needs username/password at that point.

Components of this environment:

• NetScaler Access Gateway
• Storefront
• XenApp

I anticipate it would be the access gateway providing the mutlifactor aspect.

In trying to lock down access to a web application too only users within a XenApp environment. Is it possible to send the clientname in the HTTP payload by customizing something in the XenApp server (registry, etc)? These users will be using Internet Explorer in the xenapp session.

Similar in theory to how a proxy server passes a X-Forwarded-For header.

We know WHO the end-user, since the have authenticated, but we also need to identify the end-user device distinctly from other users.

From what I've found this error is usually caused by Reporting Services, but Reporting Services is not installed on this server. Any ideas why net stop http doesn't fully work?

C:\Users\Administrator>net stop http
The following services are dependent on the HTTP service.
Stopping the HTTP service will also stop these services.

   Web Management Service
   World Wide Web Publishing Service
   Print Spooler
   Web Deployment Agent Service
   Function Discovery Resource Publication

Do you want to continue this operation? (Y/N) [N]: y
The Web Management Service service is stopping.
The Web Management Service service was stopped successfully.

The World Wide Web Publishing Service service is stopping.
The World Wide Web Publishing Service service was stopped successfully.

The Print Spooler service is stopping.
The Print Spooler service was stopped successfully.

The Web Deployment Agent Service service is stopping.
The Web Deployment Agent Service service was stopped successfully.

The Function Discovery Resource Publication service is stopping.
The Function Discovery Resource Publication service was stopped successfully.

The HTTP service is stopping........
The HTTP service could not be stopped.

I'm trying to instrument ASP.NET cache behavior on a web cluster by accessing the counters in the WMI class, Win32_PerfFormattedData_ASPNET_ASPNETApplications. 2 of my servers have this class, but 1 does not. They all have the versioned class (Win32_PerfFormattedData_ASPNET4030319_ASPNETAppsv4030319), but not the general.

How can I get WMI to expose the Win32_PerfFormattedData_ASPNET_ASPNETApplications class. It is visible in perfmon.

I have a Windows 2008 R2 64 bit virtual machine on ESXi 4.1 host. The host reports that the virtual machine is actively using less than 1 GB of memory. But, in Windows it's reporting the machine is using 7 GB of memory, even though the total of the processes listed in task manager is less than 1 GB. The machine is rather unresponsive and I'm concerned this is impacting other applications (server's purpose is to run ASP.NET state server process, which has been having trouble and led me to spot the memory question).

I just noticed High memory usage Windows Server 2008r2 on VMware and will be looking through those documents more, but what is causing this?

I have a virtual machine (Windows 2008 Server R2 64 bit running on vmware esxi 4) with 6 GB of installed RAM. It's reporting 95% memory usage and close to 0% CPU. This server hosts some WCF services on IIS and has AppFabric installed which installs its own SQL Server 2008 R2 Express. All in all, pretty low usage though.

When looking at the memory tab of the resource monitor, there is no obvious process consuming very much memory. The sum of the commit column for all processes listed is around 1.5 GB. So, while the machine is reporting 95% usage (over 5 GB), I can't figure out what's using it.

Looking at perform, I added several counters in the Process object and the biggest anomaly I found was that sqlserver had a virtual bytes count greater than 6 GB, yet it's private bytes count is 125 MB.

Does that high virtual byte count prove sql is the source of the high memory use? What would cause that spike? Or should I be looking at something else to isolate the memory hog?

I recently migrated an app from SQL Server backed sessions to the ASP.NET State Server. About once a week there is a brief failure of ASP.NET to connect to the state server. These last from 20 seconds to 2 minutes.

Unable to make the session state request to the session state server. Please ensure that the ASP.NET State service is started and that the client and server ports are the same.  If the server is on a remote machine, ....

I've been running a data collector to identify possible causes and have isolated data for the outage windows, attached.

https://dl.dropbox.com/u/3650353/session-performance.xlsx

The one main concern is that during the outages, the % CPU for the asp_state.exe process jumps significantly higher. I had theorized mcaffee was a culprit, but there are other times when mcshield.exe spikes and there is no corresponding session outage.

I bumped the timeout from 10 to 15 seconds this week as well, but just had an outage today.

Does this data point to a culprit? What could the State Server be doing during these CPU spikes? Any other counters I should be tracking?

Working with security in SQL Server 2008, I'm trying to grant a user CREATE TABLE permissions, but only within a specific schema. Does CREATE TABLE permission only apply at the database level?

Can I limit the user to create table only within a schema?

I've tried:

USE [databasename]
GRANT CONTROL ON Schema :: [schemaname] TO [username]
GO

and

USE [databasename]
GRANT ALTER ON Schema :: [schemaname] TO [username]
GO

But, the user is still unable to create a table within the target schema. It's not until I run this, hat the user can create a table:

USE [databasename]
GRANT CREATE TABLE to [username]
GO

GRANT

GRANT Database Permissions

I'm running into some problems with a file share used by an ASP.NET web application. With this configuration, there are 2 web servers (win2k8 web) that connect to a file server (win2k8 enterprise), reading and writing files using a file share. Recently, one of the web servers has begun encountering an error accessing the file share:

 IOException: The specified network name is no longer available.

There does not appear to be much info on the web for explaining what's causing this and how to best fix it, so I'm looking at what I can monitor in order to get clues. I'm not sure if it's hardware, just a load issue, file size, frequency, etc.

With Windows perfmon, what can I monitor on the File Server side? There's the "Files Open" object, any other good ones?

What can I monitor on the web server side?

EDIT:

I'll add that the UNC path uses the IP address of the file server, not a name to resolve. Also the share is a single, flat directory with over 100K files.

I have a Windows Server at a data center and an EC2 Windows Server running in the cloud. In the *Nix world I know the answer to sync these would be rsync. What are some good options for Windows? Is FTP over SSL the only choice here? What programs are good for syncing thousands of files over FTP? Robocopy?

I'm working on estimating how long it would take to download 100 GB worth of files from Amazon S3 to an EBS drive. Assuming both are in the same region, what kind of speed could be expected?

I'm looking at implementing a secure off-site data backup of about 100 GB (and growing), more than likely on Amazon S3. I haven't found many reviews of DiskAgent, has anybody had a good experience with it? Seems to be significantly more expensive than Jungle Disk, so I'm curious to know why and if it's worth it.

Can either of these run syncing processes as a service? Or are both GUI based?

With Windows Vista Business, I'd like to completely disable any group policy (both local and domain) control over the Windows Firewall so that I can turn it off and on at will on the computer directly. I have admin rights on the computer and the domain and I thought I removed all the applicable group rules controlling the firewall, but when I open the Firewall settings it still tells me that group policy is managing it and won't let me turn it on or off.

Is there a way to tell which group policy, Local or Domain is in effect?

This is my first experience with group policy, so any guidance on freeing up this control is appreciated.

Possible Duplicate:
Can you help me with my software licensing question?

I'm currently working on a high availability solution for SQL Server 2008. I have 2 servers with 1 license of standard edition SQL Server. Can you implement log shipping with just 1 license? For example if you had SQL installed on the 2nd box, but not running, and the primary server ship logs to the backup?