AWX is installed on CentOS machine. A separate machine (gitlab.techraf) serves GitLab's web interface over HTTPS with a certificate signed by a private CA.

The storyline:

1. On a "fresh" system I issue the command:

   /usr/bin/git clone --origin origin https://gitlab.techraf/techraf/project.git /tmp/project
   

   and get (expected):

   fatal: unable to access 'https://gitlab.techraf/techraf/project.git/': Peer's Certificate issuer is not recognized.

2. I install the root CA certificate on the AWX machine:

   update-ca-trust force-enable
   curl https://certificate.source/certificates/techrafCA.pem > /etc/pki/ca-trust/source/anchors/techrafCA.pem
   update-ca-trust extract
   

3. I retry the above git clone, this time getting a correct response:

   Cloning into '/tmp/project'...
   remote: Counting objects: 3, done.
   remote: Total 3 (delta 0), reused 0 (delta 0)
   Unpacking objects: 100% (3/3), done.

   ^{As a sanity check I ran git clone using different user accounts - no problem after root CA certificate installation in point 2 - the root CA seems to be installed system-wise. curl also accepts the gitlab.techraf's certificate.}

4. (problem) I define a project in AWX interface and try to synchronise it from the GitLab, however I get:

   TASK [update project using git]
   ************************************************
   fatal: [localhost]: FAILED! => {"changed": false, "cmd": "/usr/bin/git clone --origin origin https://gitlab.techraf/techraf/project.git /var/lib/awx/projects/_6__project", "failed": true, "msg": "fatal: unable to access 'https://gitlab.techraf/techraf/project.git/': Peer's Certificate issuer is not recognized.", "rc": 128, "stderr": "fatal: unable to access 'https://gitlab.techraf/techraf/project.git/': Peer's Certificate issuer is not recognized.\n", "stderr_lines": ["fatal: unable to access 'https://gitlab.techraf/techraf/project.git/': Peer's Certificate issuer is not recognized."], "stdout": "Cloning into '/var/lib/awx/projects/_6__project'...\n", "stdout_lines": ["Cloning into '/var/lib/awx/projects/_6__project'..."]}

   ^{As a sanity check I tried with a project from GitHub and AWX fetches it correctly.}

   ^{As another sanity check I ran ansible localhost -m command -a "/usr/bin/git clone --origin origin https://gitlab.techraf/techraf/project.git /tmp/project2, as well as from a different machine to ensure non-interactive shell sessions work the same. In either case the certificate was accepted.}

What can be the cause of git (called by AWX/Ansible) not using the installed techrafCA.pem certificate?

What is the next step I can take to troubleshoot the issue?

I am trying to deploy a CoreOS machine from OVA on VMware ESXi 6.0 web client (not vCenter).

I want to customise the machine before first boot, however the "Power on after deployment" checkbox seems to be missing in this interface:

After clicking Finish the machine starts up. How can I prevent the newly deployed machine from starting?

When trying to delete Azure Resource Group in the Azure Portal I get an error pointing to a Storage Account which could not be deleted.

Indeed, trying to delete the Storage Account results in an error:

Failed to delete storage account 'clie6aa7aa53aa7ace414833'. Error: The storage account cannot be deleted due to its artifacts being in use. For more information on troubleshooting this issue, see https://azure.microsoft.com/documentation/articles/storage-cannot-delete-storage-account-container-vhd/

The article is not helpful, I have already deleted all blobs and files. There are no remaining containers on the SA:

When trying to delete from PowerShell I get:

PS C:\> Remove-AzureRmStorageAccount -ResourceGroupName myresourcegroupname001 -StorageAccountName cl
i6c4ccaf5d718f00c14833

Confirm
Remove Storage Account 'cli6c4ccaf5d718f00c14833' and all content in it
[Y] Yes  [N] No  [S] Suspend  [?] Help (default is "Y"): y
Remove-AzureRmStorageAccount : Operation returned an invalid status code 'Conflict'
At line:1 char:1
+ Remove-AzureRmStorageAccount -ResourceGroupName myresourcegroupname00 ...
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : CloseError: (:) [Remove-AzureRmStorageAccount], CloudException
    + FullyQualifiedErrorId : Microsoft.Azure.Commands.Management.Storage.RemoveAzureStorageAccountCommand

Currently I have two such undeletable, empty storage accounts - one of which remains for over 72 hours.

Is there anything I can do to delete this Storage Account?

I have a basic cloud-config.yaml to run four containers on CoreOS (directly, without clustering). Two of the containers (nginx-gen and nginx-letsencrypt) mount volumes from the container named nginx. Each container startup is defined as a systemd unit.

I have added After= and Requires= dependencies, however upon first login I am greeted with:

CoreOS stable (1185.5.0)
Failed Units: 2
  letsencrypt.service
  nginx-gen.service

nginx (the one providing volume) is running.

If after that I execute:

sudo systemctl start nginx-gen.service
sudo systemctl start letsencrypt.service

They start and run correctly, so I guess the problem is in dependencies at the system startup.

I am not experienced with CoreOS nor systemd, so the cause may be pretty basic (like the unit getting "complete"-status before Docker container actually runs).

And looking at the journalctl log, indeed the nginx-gen service fails before it has a chance to mount volumes from nginx:

...
Jan 02 14:35:47 core-01 systemd[1]: nginx-gen.service: Failed with result 'start-limit-hit'.
...
Jan 02 14:36:03 core-01 docker[1401]: Status: Downloaded newer image for nginx:latest
...

What can I do to fix the problem?

My cloud-config.yaml:

#cloud-config

write_files:  
  # omitted for ServerFault: ensure template file
  #   https://raw.githubusercontent.com/jwilder/nginx-proxy/master/nginx.tmpl)
  #   and directories (/path/to/templates, /path/to/certs) are in-place

coreos:
  units:
    - name: nginx.service
      command: start
      content: |
        [Unit]
        Description=Generic Nginx container

        [Service]
        Restart=always
        ExecStart=/usr/bin/docker run --name nginx -p 80:80 -p 443:443 -v /etc/nginx/conf.d -v /etc/nginx/vhost.d -v /usr/share/nginx/html -v /path/to/certs:/etc/nginx/certs:ro nginx
        ExecStop=/usr/bin/docker rm -f nginx

    - name: nginx-gen.service
      command: start
      content: |
        [Unit]
        Description=Docker-gen service for Nginx

        After=nginx.service
        Requires=nginx.service

        [Service]
        Restart=always
        ExecStart=/usr/bin/docker run --name nginx-gen --volumes-from nginx -v /path/to/templates:/etc/docker-gen/templates -v /path/to/certs:/etc/nginx/certs:ro -v /var/run/docker.sock:/tmp/docker.sock:ro jwilder/docker-gen -notify-sighup nginx -watch -wait 5s:30s /etc/docker-gen/templates/nginx.tmpl /etc/nginx/conf.d/default.conf
        ExecStop=/usr/bin/docker rm -f nginx-gen

    - name: letsencrypt.service
      command: start
      content: |
        [Unit]
        Description=Ngnix reverse proxy Letsencrypt companion

        After=nginx.service
        Requires=nginx.service

        [Service]
        Restart=always
        ExecStart=/usr/bin/docker run --name nginx-letsencrypt -e "NGINX_DOCKER_GEN_CONTAINER=nginx-gen" --volumes-from nginx -v /path/to/certs:/etc/nginx/certs:rw -v /var/run/docker.sock:/var/run/docker.sock:ro jrcs/letsencrypt-nginx-proxy-companion
        ExecStop=/usr/bin/docker rm -f nginx-letsencrypt

Every now and then I am facing a situation where some old DSC Node remains on the registration list in the Azure Automation DSC pull server.

The nodes appear for a brief moment (1-2 seconds) after clicking on the DSC Nodes menu item (or tile) and disappear. I cannot access them otherwise. The number of DSC Nodes also includes those registrations.

The virtual machine on the screenshots does not exist anymore, adding a new one with the same name and in the same Resource Group creates a new registration entry.

How can I remove the old registrations (other than deleting and recreating the Automation Account)?

All information I could find regarding DSC on Nano Server suggest the DSC should be added to the server image upon creation, e.g. Using DSC on Nano Server:

DSC on Nano Server is an optional package in the NanoServer\Packages folder of the Windows Server 2016 media. The package can be installed when you create a VHD for a Nano Server by specifying Microsoft-NanoServer-DSC-Package as the value of the Packages parameter of the New-NanoServerImage function.

Is there any way to add the DSC to an already existing system? Namely to the one deployed from 2016-Nano-Server image provided by Microsoft on Azure.

I am trying to use cChoco module to provision a few packages with Azure Automation DSC on Windows Server 2016 virtual machines (SKU: 2016-Datacenter).

• Some packages (e.g. vcredist2008, vcredist2010)

  • Chocolatey (called from DSC) installs them successfully
  • Chocolatey (called from CustomScriptExtension) installs them successfully
  • Chocolatey (called from an interactive command prompt) installs them successfully

• Other packages (e.g. mssqlserver2014express and mssqlservermanagementstudio2014express)

  • Chocolatey (called from DSC)

    • fails to install them

    • the direct cause seems to be indicated by the error message "The system cannot find the file specified" for C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio\setup.exe;

  • Chocolatey (called from CustomScriptExtension)

    • fails to install them

    • again, the direct cause seems to be indicated by the error message "The system cannot find the file specified" for the same file as above;

  • Chocolatey (called from an interactive command prompt)

    • installs them successfully

The last action suggests a difference between and interactive session and the environment of DSC (or CustomScriptExtension), however the error message itself doesn't clearly indicate the direct reason.

Question:

What exactly could be the reason for these MSSQL Express-related packages (not all packages) to fail?

Is there any way (e.g. parameter, configuration setting) to enable these packages to be installed?

Details:

For the CustomScriptExtension (fail) and interactive command prompt (success), I used the following call (from the PowerShell script):

choco install -y mssqlservermanagementstudio2014express

For Azure Automation DSC, I have the following tasks in the configuration file (an excerpt):

cChocoPackageInstaller installVCRedist201WithChoco
{
    Name                 = 'vcredist2010'
    Ensure               = 'Present'
    DependsOn            = '[cChocoInstaller]installChoco'
}

cChocoPackageInstaller installmssqlservermanagementstudio2014express
{
    Name                 = 'mssqlservermanagementstudio2014express'
    Ensure               = 'Present'
    DependsOn            = '[cChocoInstaller]installChoco'
}

The machine is clean and vcredist2010 installs fine, however mssqlservermanagementstudio2014express fails with the following error:

2016-11-22 08:19:34,256 [INFO ] - Downloading MsSqlServerManagementStudio2014Express 64 bit
  from 'https://download.microsoft.com/download/2/A/5/2A5260C3-4143-47D8-9823-E91BB0121F94/SQLManagementStudio_x64_ENU.exe'
2016-11-22 08:20:43,755 [INFO ] -
2016-11-22 08:20:44,986 [INFO ] - Download of SQLManagementStudio.exe (850.63 MB) completed.
2016-11-22 08:22:54,657 [INFO ] - C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio.exe
2016-11-22 08:22:54,922 [INFO ] - Extracting...
2016-11-22 08:23:01,164 [INFO ] - Installing...
2016-11-22 08:23:01,745 [INFO ] - Installing MsSqlServerManagementStudio2014Express...
2016-11-22 08:23:03,311 [WARN ] - WARNING: Unable to generate 'C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio\setup.exe.ignore'
2016-11-22 08:23:03,999 [WARN ] - WARNING: May not be able to find 'C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio\setup.exe'. Please use full path for executables.
2016-11-22 08:23:05,250 [ERROR] - ERROR: Exception calling "Start" with "0" argument(s): "The system cannot find the file specified"
2016-11-22 08:23:07,903 [ERROR] - The install of mssqlservermanagementstudio2014express was NOT successful.
2016-11-22 08:23:07,950 [ERROR] - Error while running 'c:\choco\lib\MsSqlServerManagementStudio2014Express\tools\ChocolateyInstall.ps1'.
 See log for details.
2016-11-22 08:23:10,208 [WARN ] -
Chocolatey installed 1/2 packages. 1 packages failed.
 See the log for details (c:\choco\logs\chocolatey.log).
2016-11-22 08:23:10,275 [INFO ] -
2016-11-22 08:23:10,321 [ERROR] - Failures
2016-11-22 08:23:10,364 [ERROR] -  - mssqlservermanagementstudio2014express (exited -1) - Error while running 'c:\choco\lib\MsSqlServerManagementStudio2014Express\tools\ChocolateyInstall.ps1'.
 See log for details.

The log file contains:

2016-11-22 08:23:01,164 [INFO ] - Installing...
2016-11-22 08:23:01,665 [DEBUG] - Running Install-ChocolateyInstallPackage -validExitCodes '0 3010' -packageName 'MsSqlServerManagementStudio2014Express' -fileType 'EXE' -silentArgs '/IACCEPTSQLSERVERLICENSETERMS /Q /ACTION=install /UPDATEENABLED=FALSE' -file 'C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio\setup.exe' 
2016-11-22 08:23:01,745 [INFO ] - Installing MsSqlServerManagementStudio2014Express...
2016-11-22 08:23:03,311 [WARN ] - WARNING: Unable to generate 'C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio\setup.exe.ignore'
2016-11-22 08:23:03,734 [DEBUG] - Running Start-ChocolateyProcessAsAdmin -validExitCodes '0 3010' -workingDirectory 'C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio' -statements '/IACCEPTSQLSERVERLICENSETERMS /Q /ACTION=install /UPDATEENABLED=FALSE ' -exeToRun 'C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio\setup.exe' 
2016-11-22 08:23:03,953 [DEBUG] - Elevating Permissions and running ["C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio\setup.exe" /IACCEPTSQLSERVERLICENSETERMS /Q /ACTION=install /UPDATEENABLED=FALSE ]. This may take a while, depending on the statements.
2016-11-22 08:23:03,999 [WARN ] - WARNING: May not be able to find 'C:\Windows\TEMP\chocolatey\chocolatey\MsSqlServerManagementStudio2014Express\SQLManagementStudio\setup.exe'. Please use full path for executables.
2016-11-22 08:23:04,781 [DEBUG] - Setting RunAs for elevation
2016-11-22 08:23:05,250 [ERROR] - ERROR: Exception calling "Start" with "0" argument(s): "The system cannot find the file specified"
2016-11-22 08:23:05,509 [DEBUG] - Built-in PowerShell host called with ['[System.Threading.Thread]::CurrentThread.CurrentCulture = '';[System.Threading.Thread]::CurrentThread.CurrentUICulture = ''; & import-module -name 'c:\choco\helpers\chocolateyInstaller.psm1'; & 'c:\choco\helpers\chocolateyScriptRunner.ps1' -packageScript 'c:\choco\lib\MsSqlServerManagementStudio2014Express\tools\ChocolateyInstall.ps1' -installArguments '' -packageParameters '''] exited with '-1'.
2016-11-22 08:23:05,557 [DEBUG] - Calling command ['"C:\Windows\System32\shutdown.exe" /a']
2016-11-22 08:23:07,369 [DEBUG] - Command ['"C:\Windows\System32\shutdown.exe" /a'] exited with '1116'
2016-11-22 08:23:07,385 [DEBUG] - Capturing package files in 'c:\choco\lib\MsSqlServerManagementStudio2014Express'
2016-11-22 08:23:07,432 [DEBUG] -  Found 'c:\choco\lib\MsSqlServerManagementStudio2014Express\MsSqlServerManagementStudio2014Express.nupkg'
  with checksum 'AD996F040E8828374737D274BC5EF15D'
2016-11-22 08:23:07,573 [DEBUG] -  Found 'c:\choco\lib\MsSqlServerManagementStudio2014Express\MsSqlServerManagementStudio2014Express.nuspec'
  with checksum 'D7BE143BDF795902EAEE640DFCE85D15'
2016-11-22 08:23:07,619 [DEBUG] -  Found 'c:\choco\lib\MsSqlServerManagementStudio2014Express\tools\ChocolateyInstall.ps1'
  with checksum 'E60A57F683AB9A3A4D9C28ED78E09081'
2016-11-22 08:23:07,650 [DEBUG] -  Found 'c:\choco\lib\MsSqlServerManagementStudio2014Express\tools\ChocolateyUninstall.ps1'
  with checksum '4F30E08CD2C8D4E13C3AE84AA94FBE00'
2016-11-22 08:23:07,682 [DEBUG] - Attempting to create directory "c:\choco\.chocolatey\MsSqlServerManagementStudio2014Express.12.2.5000.20161110".
2016-11-22 08:23:07,749 [DEBUG] - Attempting to copy "c:\choco\.chocolatey\MsSqlServerManagementStudio2014Express.12.2.5000.20161110\.files.update"
 to "c:\choco\.chocolatey\MsSqlServerManagementStudio2014Express.12.2.5000.20161110\.files".
2016-11-22 08:23:07,838 [DEBUG] - Attempting to delete file "c:\choco\.chocolatey\MsSqlServerManagementStudio2014Express.12.2.5000.20161110\.files.update".
2016-11-22 08:23:07,854 [DEBUG] - Attempting to delete file "c:\choco\.chocolatey\MsSqlServerManagementStudio2014Express.12.2.5000.20161110\.sxs".
2016-11-22 08:23:07,871 [DEBUG] - Attempting to delete file "c:\choco\.chocolatey\MsSqlServerManagementStudio2014Express.12.2.5000.20161110\.pin".
2016-11-22 08:23:07,885 [DEBUG] - Sending message 'HandlePackageResultCompletedMessage' out if there are subscribers...
2016-11-22 08:23:07,903 [ERROR] - The install of mssqlservermanagementstudio2014express was NOT successful.
2016-11-22 08:23:07,950 [ERROR] - Error while running 'c:\choco\lib\MsSqlServerManagementStudio2014Express\tools\ChocolateyInstall.ps1'.
 See log for details.
2016-11-22 08:23:07,994 [DEBUG] - Moving 'c:\choco\lib\MsSqlServerManagementStudio2014Express'
 to 'c:\choco\lib-bad\MsSqlServerManagementStudio2014Express'
2016-11-22 08:23:10,053 [DEBUG] - Attempting to delete file "C:\Windows\system32\config\systemprofile\AppData\Local\NuGet\Cache\MsSqlServerManagementStudio2014Express.12.2.5000.20161110.nupkg".
2016-11-22 08:23:10,208 [WARN ] - 
Chocolatey installed 1/2 packages. 1 packages failed.
 See the log for details (c:\choco\logs\chocolatey.log).
2016-11-22 08:23:10,275 [INFO ] - 
2016-11-22 08:23:10,321 [ERROR] - Failures
2016-11-22 08:23:10,364 [ERROR] -  - mssqlservermanagementstudio2014express (exited -1) - Error while running 'c:\choco\lib\MsSqlServerManagementStudio2014Express\tools\ChocolateyInstall.ps1'.
 See log for details.
2016-11-22 08:23:10,396 [DEBUG] - Sending message 'PostRunMessage' out if there are subscribers...
2016-11-22 08:23:10,431 [DEBUG] - Exiting with -1

I am trying to get a remote PowerShell session to a Windows Server 2016 deployed on Azure via its public IP address (initially I was trying to configure Server management tools gateway, however I can't get it working even between two identical, freshly deployed servers).

I have two servers (target 10.0.0.4 and source 10.0.0.5) connected to the same subnet and I can connect from the source to the target on local network (10.0.0.0):

PS C:\> Set-Item WSMan:\localhost\Client\TrustedHosts -Value "10.0.0.4" -Concatenate -Force
PS C:\> Enter-PSSession -ComputerName 10.0.0.4 -Credential (Get-Credential)

cmdlet Get-Credential at command pipeline position 1
Supply values for the following parameters:
Credential
[10.0.0.4]: PS C:\Users\techraf\Documents> exit

If I however try to use the target's public IP address the connection gets refused:

PS C:\> Set-Item WSMan:\localhost\Client\TrustedHosts -Value "13.85.78.151" -Concatenate -Force
PS C:\> Enter-PSSession -ComputerName 13.85.78.151 -Credential (Get-Credential)

cmdlet Get-Credential at command pipeline position 1
Supply values for the following parameters:
Credential
Enter-PSSession : Connecting to remote server 13.85.78.151 failed with the following error message : WinRM cannot
complete the operation. Verify that the specified computer name is valid, that the computer is accessible over the
network, and that a firewall exception for the WinRM service is enabled and allows access from this computer. By
default, the WinRM firewall exception for public profiles limits access to remote computers within the same local
subnet. For more information, see the about_Remote_Troubleshooting Help topic.
At line:1 char:1
+ Enter-PSSession -ComputerName 13.85.78.151 -Credential (Get-Credentia ...
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : InvalidArgument: (13.85.78.151:String) [Enter-PSSession], PSRemotingTransportException
    + FullyQualifiedErrorId : CreateRemoteRunspaceFailed

Prior to trying I have:

• created an inbound rule in the NSG for TCP ports 5985 and 5986 (for the whole internet)
• disabled (completely, for a test) Windows firewall on the target system

I also checked the target ports with nmap from the source machine, and while 5985 is open, 5986 is closed (including connection on 10.0.0.0 network).

Is there anything else required to make it work?

When deploying a template through Azure Portal, i.e. by following the link:

https://portal.azure.com/#create/Microsoft.Template/uri/https%3A%2F%2Fraw.githubusercontent.com%2FAzure%2Fazure-quickstart-templates%2Fmaster%2F{template name}%2Fazuredeploy.json

All deployments created this wah are named Microsoft.Template (and subsequent ones Microsoft.Template_1, Microsoft.Template_2, etc.) in the Deployments section of the portal:

Is it possible to customize the name (either by modifying the template, or provide a different name when deploying)?

Guessing, I tried adding name at the root level of the azuredeploy.json, but it gets refused with a wrong syntax message.

I would like to deploy a template from the Azure Resource Manager QuickStart Templates.

Many of the templates there have a default value specified in the template file (for example this one azuredeploy.json ):

"ouPath": {
  "type": "string",
  "defaultValue": "",
  "metadata": {
    "description": "Specifies an organizational unit (OU) for the domain account. Enter the full distinguished name of the OU in quotation marks. Example: 'OU=testOU; DC=domain; DC=Domain; DC=com"
  }
},

However, the corresponding azuredeploy.parameters.json does not define the ouPath value.

When I specify the file in --parameters-file in the Azure CLI call:

azure group deployment create "${resource_group_name}" "${deployment_name}" --template-uri "${template_uri}" --parameters-file ${parameters_file}

I get an error:

info: Executing command group deployment create
  error: Template and Deployment "parameters" objects lengths do not match
  Deployment Parameter file does not have { ouPath } defined.
  error: Error information has been recorded to /Users/techraf/.azure/azure.err
  error: group deployment create command failed

If I define add the value explicitly to the parameters.json file, the deployment works properly.

Main question:

The sheer number of examples (parameters.json) in the above Azure repository without explicitly specified values makes me think, there must be a way to tell Azure to use default value instead of providing it in parameters.json.

Is there a way to avoid adding the values that have defaultValue defined in the template file when using Azure CLI?

Additional question:

Worse yet, in the example I have given above, the default value is empty and if I specify an empty value in the parameters.json:

"ouPath": {
  "value": ""
},

I get an interactive prompt for the value.

How can I avoid that?

I have the following check in Serverspec tests:

openssl_latest_version = 'OpenSSL 1.0.2h  3 May 2016'

describe command('openssl version') do
  its(:stdout) { should match openssl_latest_version }
end

Is there a way to automatically fetch the string that the latest officially released OpenSSL returns on executing openssl version? I.e. automatically set openssl_latest_version based for example on some curl call.

My intention is that the test should fail if installed version is later than the one currently published by OpenSSL as latest at the moment of running the test, but without human updating the parameters.

I do not want to check it against a version available in package manager for any specific OS version, I do not want to install latest version, I simply want the test to fail when the running version is different than the one announced as recent.

I am using a multi-hop SSH connection with agent forwarding enabled at each hop. I want to connect to a destination machine (Machine D) using the key of the preceding hop, not the one from the originating machine.

The details:

Working on the console of Machine A I connect to Machine B with SSH (ForwardAgent yes):

Machine A  ---SSH(id_rsa_A)--->  Machine B  ---SSH(forwarded_key)--->  Machine D1

When I execute ssh-add -L (on B) I get the id_rsa_A. And if I connect from B to D1, it uses id_rsa_A for authentication.

OK, this is what I want.

Working on the console of Machine B I connect to Machine C with SSH (ForwardAgent yes):

Machine B  ---SSH(id_rsa_B) --->  Machine C  ---SSH(forwarded_key)--->  Machine D2

When I execute ssh-add -L (on C) I get the id_rsa_B. And if I connect from C to D2, it uses id_rsa_B for authentication.

OK, this is what I want.

Now I go back to Machine A and have a two-hop SSH connection both with ForwardAgent yes to Machine C via B:

Machine A  ---SSH(id_rsa_A)--->  Machine B  ---SSH(id_rsa_B)--->  Machine C ---SSH(forwarded_key)--->  Machine D2

When I execute ssh-add -L (on C) I get the id_rsa_A. And when I connect to Machine D2, SSH connection uses id_rsa_A even though the preceding connection between B and C explicitly uses id_rsa_B.

This is what I want to change. Namely I want forwarded_key to be id_rsa_B.

Is it possible (and how) to force Machine C (connected from A via B) to use id_rsa_B for its outgoing SSH connections?

I even tried ForwardAgent no option on first connection:

Machine A  ---SSH(id_rsa_A)--->  Machine B

leaving ForwardAgent yes on subsequent connection:

Machine B  ---SSH(id_rsa_B)--->  Machine C

but then further connection does not use agent forwarding at all:

Machine C ---SSH(no_key)--->  Machine D2

It looks like it's only for the whole pipe between console and the currently connected machine.

I created a simple Ansible role with the purpose of configuring dnsmasq on OSX targets (dnsmasq is installed from homebrew).

I want to create a handler that would restart the daemon when configuration file gets changed (Ansible as of version 1.9.3 does not implement service module for OSX).

I have a task:

- name: Create dnsmasq config file in /usr/local/etc/ from template
  template: src=dnsmasq.conf.j2 dest=/usr/local/etc/dnsmasq.conf
  notify:
  - restart dnsmasq

In /Library/LaunchDaemons/homebrew.mxcl.dnsmasq.plist there is a key <key>KeepAlive</key> which starts the daemon automatically when not running, so the following handler does its job. Can I always rely on this?

- name: restart dnsmasq
  sudo: yes
  command: launchctl stop homebrew.mxcl.dnsmasq

Aren't there any scenarios or timing issues that would result in the daemon getting permanently stopped? What would be the proper way for such handler to be written?