I’m trying to use web deploy to deploy my dacpac package, which comes to executing some sql scripts.
I have local windows account called .\DeploymentService, which is in local Administrators group, which I want to own the database and execute scripts.
For that - I configure delegation accordingly - In Management Service Delegation I set "Specific User" for dbDacFx rule = .\DeploymentService providing password
I create according serveradmin login in SQL Server. My WMSvc executed under LOCAL SERVICE account.
I use the following command line parameters for deployment:
msdeploy.exe
-verb:sync
-source:dbDacFx="C:\Main\Src\Community.DB\bin\Debug\Community.DB.dacpac"
-dest:dbDacFx="Data Source=.;Database=CommunityInt; Integrated Security=true",computername=”https://Community02:8172/msdeploy.axd?site=Default
Web Site”,username=.\DeploymentService,password=*************,authType=basic
-allowuntrusted
I execute it on my PC, where destination is different PC.
However this fails with error “Invalid Handle” or “Class name not found” depending on do I have “Local service” login with public role in my SQL Server created.
Expected behavior:
When I set user name in Management Service Delegation to specific account, I expect MSDeploy to be executed under the account I specified.
Actual behavior:
I traced using SQL profiler in target environment and I found out that WMSvc executes msdeploy under its process account (LOCAL SERVICE) instead of .\DeploymentService, and that’s why script execution fails. If in SQL server I have LOCAL SERVICE account mapped to serveradmin role, then it works fine. If I execute WMSvc under .\DeploymentService account, it also works fine.
So basically there is NO WAY TO USE "User Name" in Management Service Delegation - It just does not matter what you set up there - it gives no effect.
Does any one know how to make that work?
Keywords: WebDeploy, WMSvc, dbDacFx, Impersonation, Delegation
Hey guys I'm sorry to hear that you are running into this issue. I wanted to let you know that we have a bug in the dbDacFx provider/MSDeploy which is preventing SQL Auth to work when used with the dbDacFx provider in WMSvc scenarios.
We have not yet received enough feedback regarding this to warrant servicing MSDeploy in order to unblock this. If you are impacted by this the best thing to do is to create an entry at http://aspnet.uservoice.com and vote it up. If we get enough votes then we can consider shipping an update to unblock this. Sorry for the bad news.
Related
I'm attempting to query a remote machine with the PowerShell cmdlet: Get-WmiObject and using credentials for a user in Azure AD that has the Azure AD joined device local administrator role and the IAM role assignment of Virtual Machine Administrator Login on the remote machine.
When I run the following command:
Get-WmiObject -Class Win32_Process -Namespace "root/cimv2" -ComputerName <remote_computer_local_ip> -Impersonation Impersonate -Credential AzureAD\<username>
I receive the following error message:
Get-WmiObject : A security package specific error occurred. (Exception from HRESULT: 0x80070721)
Is the ability to run remote WMI queries supported when using an Azure AD user?
Remote machine specs:
Windows Server 2019 Datacenter
10.0.17763 Build 17763
Other troubleshooting notes:
I'm able to sign to the remote machine with the same user and have followed the steps in the https://learn.microsoft.com/en-us/azure/active-directory/devices/howto-vm-sign-in-azure-ad-windows document
When running the same command on the remote machine locally it properly returns the Win32_Process data
If I use the local administrator credentials on the remote command it also works which suggests that remote WMI is working
Both computers are Azure AD joined and show AzureAdJoined : YES from the dsregcmd /status command
I've attempted to use the Computer Name, Local IP, and FQDN and they all yield the same result
After talking with Microsoft support about this issue, it seems that the error stems from the different authentication methods between a local user account and an Azure AD account.
It seems it is not supported as they are using different authentication protocols.
Azure AD user uses OAuth and the security error is due to Kerberos.
Here is the reference article for more details: Authentication protocols in Azure Active Directory B2C | Microsoft Docs
They have also mentioned that there is no current information on whether or not this will be supported in the future and have recommend checking Azure updates for future product updates.
It could happen for number of reasons. Please check if any of these in your case is the one.
When You try to setup a WMI session when:
a. There are more than one computer accounts with the same name in play and one of those stale computer accounts resides in the same domain as your user account
b. Or when servers operating reside in the same AD DS forest but in different domains.
When server is not able to find the user object due to password credentials that may be stored as cache from previous login.
Incorrect time on machines/server that doesn’t match.
If firewalls are enabled for server which may restrict the access.
possible work arounds
Try azuread\user#mydomain.com or user#mydomain.com or domainame\username format to login or Try taking out the domain and adding it back again.
The local PC and remote PC must be in the same Azure AD tenant.
Run AAD delta Sync to make sure everything is Synced and make sure of internet connectivity and network.
Try if you can turn off firewall.
References:
remote access - AAD- Server Fault
rdp -Azure AD Joined - Server Fault
«0x80070721 | exchange12rocks.org)
A security package specific error occurred | Microsoft Docs
After a successful build using a hosted VS2017 agent, I'm trying to deploy on-prem using an local hosted agent which was setup using a domain account which should have sufficient permissions as it is in the local admin group. As I was digging around on this issue elsewhere, one user re-installed their agent using NTATHORITY\SYSTEM and it worked.
I d'ont really need to create or stop\start the website, just deploy the recent build artifact.
What permissions should I check or should I use another task?
019-04-06T21:03:10.3898646Z ERROR ( message:Configuration error
2019-04-06T21:03:10.3899503Z
2019-04-06T21:03:10.3899791Z Filename: redirection.config
2019-04-06T21:03:10.3900026Z 2019-04-06T21:03:10.3900293Z Line
Number: 0 2019-04-06T21:03:10.3900530Z 2019-04-06T21:03:10.3900852Z
Description: Cannot read configuration file due to insufficient
permissions 2019-04-06T21:03:10.3901076Z 2019-04-06T21:03:10.3901333Z
. ) 2019-04-06T21:03:10.8135484Z ##[error]Process 'appcmd.exe' exited
with code '5'.
An agent is running under a user. The error means that your user doesn't have permissions to read/modify the redirection.config file that is necessary to manage IIS.
You have multiple options to solve this.
Change the user were the agent is running under. In example: NTATHORITY\SYSTEM
Give permissions to the user were your agent is running under to the folder C:\Windows\System32\inetsrv\config. If you want to create app pools or websites etc, modify permissions are of course needed.
When we install Azure agent as a service(In Windows, search for 'Services' and look for something like Azure Pipelines Agent..), by default it is logged on as 'Network Service'. Change it to 'Local System' and the pipeline should be able to run IIS related tasks/commands.
We have a Key Vault in a resource group in an Azure instant.
We have a user in the US (ME) and a user in different country (FU).
Both of use have many things in common, namely:
Using same version of VS 2017.
Running the exact same code.
Our VS user account is the same (a user in our Azure AD instance).
We are using a Managed Identity
If I run the code in US (logged in in VS as the FU), I am able to read the secret and display on the screen.
If the FU(logged in in VS as the FU but in another country), when he runs the code it throws the following exception
Operation returned an invalid status code 'Unauthorized'
The line of code that throws the error is:
var secret = await keyVaultClient.GetSecretAsync("https://XXXXXXX.vault.azure.net/secrets/username")
.ConfigureAwait(false);
We have both installed Azure CLI 2.0.
However, I found these stipulations at this site.
Your on-premise active directory is synced with Azure AD.
You are running this code on a domain joined machine.
Neither of these are true in our case.
Possibly a good test of these would for our vendor to allow me to remote into his machine and put my identity on VS and then run the code.
If we still get the error, then it is very likely this is our problem.
The above link said we could "Run the application using a service principal in local development environment"
Would that fix the problem???
I am fairly new to Azure and C#. Any help would be greatly appreciated!
I'm sitting in a project where I will move from TFS to VSTS so we do have a working release definition.
But when I try deploying a service fabric cluster i get the following error:
2018-08-28T09:02:59.8922249Z ##[error]An error occurred attempting to acquire an Azure Active Directory token. Ensure that your service endpoint is configured properly with valid credentials. Error message: Exception calling "AcquireToken" with "3" argument(s): "AADSTS50079: Due to a configuration change made by your administrator, or because you moved to a new location, you must enroll in multi-factor authentication to access '< service principle Id >'.
Trace ID: < guid1 is here >
Correlation ID: < guid2 is here >
Then I go to the azure portal -> AAD -> Sign In -> look up my specific sign in (based on correlation Id) and there it state that Sign-in status is Success
Considering this works for our TFS instance i assume the service principle is correctly set up. But since the build/deploy agents is now on a VM in azure instead of on prem for TFS, is there anything i need to change?
Traffic should be OK, i can navigate to the https-adress to the cluster from the VM with agents.
I've tried google it, but to no success so hopefully someone can point me to the right direction where to look.
And in portal, 'MFA is required' is no, so multi factor should not be neccesary.
Just try using certificate based authentication instead of using AAD Authentication in the service endpoint configuration.
Reference the same issue here: https://github.com/Microsoft/vsts-tasks/issues/7714
If that still not work, just try to create a new endpoint, then try it again.
I have three Windows Server 2012 R2 without any AD in a DMZ network. Two servers are front end web servers with ASP.NET and one have SQL Server and a network share that both front end servers use for shared data.
My problem is how do I configure the Application Pool identity and the Network Share so the ASP.NET application can read and write to the network share?
This is simple with an AD available when you can use domain accounts for the application pool identity but there is no AD available in this setup.
I will answer my own question since I succeeded to setup the server. This is what I did:
1) Create an account with the same username and password on all three servers. Make sure that does not expire or must be changed.
2) Create a Network Share and give the new account read/write rights. I also tested that I could connect from the front end servers using the new account to verify that no firewalls are in the way.
3) Included the user in the IIS_IUSRS group that indirectly gives it Logon as Batch Job rights.
4) Run the following command to grant rights to the user
aspnet_regiis -ga <your_app_pool_user>
See more: How To: Create a Service Account for an ASP.NET 2.0 Application (MSDN)
5) Restarted WAS and IIS to make sure the changes to the accounts group membership takes hold if tried to use the account.
C:> net stop was /y
C:> net start w3svc
6) Create an Application Pool and set the Identity.
This is the part where I got stuck with error messages when trying to set the identity.
From IIS Manager I got the following error dialog: "There was an error while performing this operation. Details: Value does not fall within the expected range."
Trying to set the App Pool identity from the command line I receive a similar error:
C:> appcmd set config /section:applicationPools
/[name='test-pool'].processModel.identityType:SpecificUser
/[name='test-pool'].processModel.userName:MyAccountName
/[name='test-pool'].processModel.password:P#ssw0rd
ERROR ( hresult:80070057, message:Failed to commit configuration changes.
The parameter is incorrect.
)
When I remove the last parameter, password, the command will succeed changing identity type and setting the username but I did never figure out why I could not set the password so I retorted to editing my applicationHost.config file directly. Unfortunately with the the password ending up in clear text.
<configuration>
...
<system.applicationHost>
<applicationPools>
...
<add name="test-pool" managedRuntimeVersion="v4.0">
<processModel identityType="SpecificUser"
userName="MyAccountName" password="P#ssw0rd" />
</add>
...
</applicationPools>
...
</system.applicationHost>
...
</configuration>
7) Finally I set my Web Application to use the application and it could access the Network Share without any issues.
I had the same problem but couldn't let the password in clear text so I dig a little further and found this article :
http://social.technet.microsoft.com/wiki/contents/articles/30344.custom-iis-app-pool-identity-value-does-not-fall-within-the-expected-range.aspx
The key step to diagnose is to look at the right events :
To figure out how to resolve this, I went into the event viewer. There was nothing in the Application log, so I headed down to Applications and Services Logs => Microsoft => Windows => IIS-Configuration. The logs in here are disabled by default, so they have to be enabled. (To do so, right click the log, and choose Enable log.) Once enabled, re-run the attempt to set the identity, and refresh the view (Actions pane or F5), and voila!, now we have some more information on the error. In the results were two Errors (event ID 42 and 43).
I had the same event errors as in the article :
ID 42: Failed to initialize the 'IISWASOnlyAesProvider' encryption
provider in
'\?\C:\windows\system32\inetsrv\config\applicationHost.config'.
Please check your configuration.
ID 43: Failed to encrypt attribute
'Microsoft.ApplicationHost.AesProtectedConfigurationProvider'.
Then I did the following :
restore an old version of the ConfigEncKey.key file (to c:\windows\System32\inetsrv\config )
replace the <configProtectedData><providers> section by an old one (in c:\windows\System32\inetsrv\config\applicationHost.config )
Then I can again set a custom identity to the application pool.
Had similar problem. Reinstalled the IIS Manager and got a new applicationHost.config
When I did the WinDiff on the new and old files I noticed that the SessionKey were different. Works now.
AesProvider and IISWASOnlyAesProvider