I have an azure web site, and an application that i have set up via VSTS to build/release to this site. I'm using the Azure app service deploy task ( https://github.com/Microsoft/azure-pipelines-tasks/blob/master/Tasks/AzureRmWebAppDeploymentV4/README.md ) and i need to be able to transform the web.config based on library variable settings.
The basic transformations are easy, i have no problem just replacing the connection strings and the app settings. There are however, more difficult settings that i need to change. for example:
<sessionState mode="InProc"/>
<sessionState mode="Custom" customProvider="RedisStateStore" cookieless="false" regenerateExpiredSessionId="true">
<providers>
<add name="RedisStateStore" applicationName="XXX" type="YYY"/>
</providers>
</sessionState>
<sessionState cookieless="false" regenerateExpiredSessionId="true" mode="Custom" customProvider="SqlSessionStateProviderAsync">
<providers>
<add name="SqlSessionStateProviderAsync" connectionStringName="SQLSessionConnectionString" type="Microsoft.AspNet.SessionState.SqlSessionStateProviderAsync, Microsoft.AspNet.SessionState.SqlSessionStateProviderAsync, Version=1.1.0.0, Culture=neutral, PublicKeyToken=31bf3856ad364e35" />
</providers>
</sessionState>
based on settings in my deployment library variables, i need to be able to configure the session state to be one of the three above, and edit the appropriate settings accordingly (the variables inside the redis provider for example can change.
Additionally i need to edit more files than just the web.config. What are good ways to do this?
Related
Hopefully I can keep this as clear and simple as possible. I'm trying to replace variables in my web.config based upon an environment. Currently I am able to replace other variables such as the one in my connection string and app settings by putting the variable in the format "#{variablename}#. This is not working for the variable in the endpoint address.
I was previously using replace tokens to accomplish this but it must now be built into azure because I don't have any extra task in the build or in the release task that's performing the replacement. Unless it's somewhere else that I'm not aware of. I have searched everything I could find and I am not seeing a reason why the value in this section isn't replace I thought it was replacing based off of name but I don't see why this value won't replace.
This replacement works with VariableName ConnectionString
<connectionStrings>
<add name="ConnectionString" connectionString="#{ConnectionString}#" providerName="System.Data.SqlClient"/>
</connectionStrings>
This doesn't work with VariableName ServiceSoapEndPoint
<system.serviceModel>
<bindings>
<basicHttpBinding>
<binding name="Service" />
</basicHttpBinding>
<customBinding>
<binding name="Service1">
<textMessageEncoding messageVersion="Soap1" />
<httpTransport />
</binding>
</customBinding>
</bindings>
<client>
<endpoint address="#{ServiceSoapEndPoint}#" binding="basicHttpBinding"
bindingConfigu`enter code here`ration="ServiceSoap" contract="Service.ServiceSoap"
name="ServiceSoapEndPoint" />
</client>
I wasn't able to find the exact solution I was looking for so I ended up doing a work around to accomplish this. I was basically trying to have environment variables so that when I would publish it would change the endpoint address based on the environment. Since I was able to get the value to change in the appsettings element I did a work around to instantiate the soap service using the value read from the appsettings value.
Using a console app that connected to a DB using EF, I was able to call Update-Database and have it connect to the DB. However, I'm having trouble when trying to migrate this over to an Azure function. A similar question was asked here, and I suspect the solution might be the same, although there's not too much detail on what that might be.
In my console app, the following relevant configuration was supplied:
<configuration>
<configSections>
<section name="entityFramework" type="System.Data.Entity.Internal.ConfigFile.EntityFrameworkSection, EntityFramework, Version=6.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089" requirePermission="false" />
<connectionStrings>
<add name="DBConnection" providerName="System.Data.SqlClient" connectionString="Server=.\SQLEXPRESS;Database=MyDB;User Id= . . ." />
</connectionStrings>
<entityFramework>
<defaultConnectionFactory type="System.Data.Entity.Infrastructure.SqlConnectionFactory, EntityFramework" />
<providers>
<provider invariantName="System.Data.SqlClient" type="System.Data.Entity.SqlServer.SqlProviderServices, EntityFramework.SqlServer" />
</providers>
</entityFramework>
In my local.settings.json, I have only the DB Connection String:
{
"IsEncrypted": false,
"Values": {
"AzureWebJobsStorage": "DefaultEndpointsProtocol=https;AccountName=...",
"AzureWebJobsDashboard": "DefaultEndpointsProtocol=https;AccountName=...",
},
"ConnectionStrings": {
"DBConnection": "Server=.\\SQLEXPRESS;Database=..."
}
}
The implication in the linked question, is that the provider name needs to be supplied. I would assume that some of the other EF config sections would also need to be supplied. As it is, when I run Update-Database with the Functions app set as start-up, I get this error:
Exception calling "SetData" with "2" argument(s): "Type 'Microsoft.VisualStudio.ProjectSystem.VS.Implementation.Package.Automation.OAProject' in assembly
Which tends to just mean that the start-up project doesn't have the correct EF configuration.
My question, therefore, is what additional configuration do I need in my local.settings.json file?
You can add the provider name to your connection string in the local.settings.json file as shown here:
{
"IsEncrypted": false,
"Values": {
"AzureWebJobsStorage": "DefaultEndpointsProtocol=https;AccountName=...",
"AzureWebJobsDashboard": "DefaultEndpointsProtocol=https;AccountName=...",
},
"ConnectionStrings": {
"DBConnection": {
"ConnectionString": "Server=.\\SQLEXPRESS;Database=...",
"ProviderName": "System.Data.SqlClient"
}
}
}
Note that I took the name "DBConnection" because this is what you had in your console application and my assumption is that in your context class, that inherits from DbContext, of your Azure Function you are still referencing the same name.
My second assumption is that you are testing you Azure Function locally on your machine because you are using SQL Express.
Pay attention that if you run your Azure Function from Azure, then you need to go to your Function App's application settings (image 1) and add your connection string to the Connection strings collection (image 2). You have to do this because the connection strings declared in your local.settings.json will not be deployed to Azure.
The related Microsoft documentation can be found here: local settings file documentation.
Image 1 - Azure Function App application settings
Image 2 - Connection strings collection
Additional info: when you locally run/debug your Azure Function, you could use the Azure Storage Emulator and avoid binding your Azure Function to a storage account hosted by Azure. You just have to replace in your local.settings.json the values assigned to AzureWebJobsStorage and AzureWebJobsDashboard by UseDevelopmentStorage=true.
I am working on VSTS release task for deploying the Windows Services Project. Unfortunately, we are not creating any Build Definition for creating drop folder.
But, my client will provide drop folder for this project, what I need is “I want to override the keys of an existing .exe.config file” at release level.
For creating the Windows Services Deploy task,I followed this Windows Services Extension
For example my drop folder looks like below:
Many thanks for this reference article and It's a very useful for changing values in config file using Power Shell commands. I have doubt in from that reference link :
For Example, If had a Code like this :
<erecruit.tasks>
<tasks>
<task name="AA" taskName="AA">
<parameters>
<param key="connectionString">Server="XXXX"</param>
</parameters>
</task>
How to change this above connectionstring value?
You can use Tokenizer task in Release Management Utility tasks extension.
Install Release Management Utility tasks extension
Add Tokenizer with XPath/Regular expressions task to release definition (Specify Source filename and Configuration Json filename)
Config file sample:
<?xml version="1.0" encoding="utf-8" ?>
<configuration>
<appSettings>
<add key="TestKey1" value="__Token1__" />
<add key="TestKey2" value="__Token2__" />
<add key="TestKey3" value="__Token3__" />
<add key="TestKey4" value="__Token4__" />
</appSettings>
<startup>
<supportedRuntime version="v4.0" sku=".NETFramework,Version=v4.5.2" />
</startup>
</configuration>
Configuration Json file (Default Environment is the environment name in release definitioin):
{
"Default Environment":{
"CustomVariables":{
"Token2":"value_from_custom2",
"Token3":"value_from_custom3"
},
"ConfigChanges":[
{
"KeyName":"/configuration/appSettings/add[#key='TestKey1']",
"Attribute":"value",
"Value":"value_from_xpath"
}
]
}
}
Then the value of TestKey1 (key) will be related to value_from_xpath and the values of TestKey2 and TestKey3 will be related to value_from_custom2 and value_from_custom3.
On the other hand, you can use release variables directly if you don’t specify Configuration Json filename.
For example, there is __TokenVariable1__ in your config file and TokenVariable1 release/environment variable in release definition, then the __TokenVariable1__ will be replaced through Tokenizer task.
A related article: Using Tokenization (Token Replacement) for Builds/Releases in vNext/TFS 2015
Update:
You also can do it through PowerShell directly.
Update configuration files using PowerShell
I recently switched from ODP Unmanaged to ODP Managed (in conjunction with Entity Framework).
The Unmanaged drivers were working fine after adding the necessary information in the web.config section. I could add the stored procedures and generate the complex types using the Function Import - Get Column information (I'm trying to import a stored procedure with an OUT refcursor parameter).
After the switch the config section was updated to reflect the new format and everything works at runtime (so the format is correct).
However when I try to generate the complex types again (or add a new Function Import) I just get a System.notSupportedException Message: The specified type is not supported by this selector) Without any indication which type/selector it is (obviously)...
Google has turned up nothing and the thread on the Oracle Forums has gathered no response as well.
Versions:
ODP.Net (ODAC) : v12.1 (Production release; DLL v4.121.1.0)
EF v5
.NET v4.5
Config file (trimmed a bit):
<configSections>
<section name="oracle.manageddataaccess.client" type="OracleInternal.Common.ODPMSectionHandler, Oracle.ManagedDataAccess"/>
</configSections>
<oracle.manageddataaccess.client>
<version number="*">
<edmMappings>
<edmMapping dataType="number">
<add name="bool" precision="1"/>
<add name="byte" precision="2" />
<add name="int16" precision="5" />
<add name="int32" precision="10" />
<add name="int64" precision="38" />
</edmMapping>
</edmMappings>
<implicitRefCursor>
<storedProcedure schema="ECOM" name="SHP_API_ORDERS.CREATE_ORDER">
<refCursor name="O_RS">
<bindInfo mode="Output"/>
<metadata columnOrdinal="0" columnName="COL1" nativeDataType="Number" providerType="Decimal" allowDBNull="false" numericPrecision="10" numericScale="0" />
<metadata columnOrdinal="1" columnName="COL2" nativeDataType="Date" providerType="Date" allowDBNull="true" />
<metadata columnOrdinal="2" columnName="COL3" nativeDataType="Varchar2" providerType="Varchar2" allowDBNull="false" columnSize="10" />
</refCursor>
</storedProcedure>
</implicitRefCursor>
</version>
</oracle.manageddataaccess.client>
<entityFramework>
<defaultConnectionFactory type="System.Data.Entity.Infrastructure.SqlConnectionFactory, EntityFramework" />
</entityFramework>
<system.data>
<DbProviderFactories>
<remove invariant="Oracle.ManagedDataAccess.Client" />
<add name="ODP.NET, Managed Driver"
invariant="Oracle.ManagedDataAccess.Client"
description="Oracle Data Provider for .NET, Managed Driver"
type="Oracle.ManagedDataAccess.Client.OracleClientFactory, Oracle.ManagedDataAccess, Version=4.121.1.0, Culture=neutral, PublicKeyToken=89b483f429c47342" />
</DbProviderFactories>
</system.data>
The implicit ref cursor config file format is different between Unmanaged ODP.NET and Managed ODP.NET. That might be part of your problem.
To save yourself from pulling your hair out, install the latest Oracle Developer Tools for Visual Studio (ODT) and use the new feature that automatically generates this config:
1) Install ODT 12.1 if you haven't already
2) Find the stored procedure in server explorer, right click it and run it, and enter input parameters.
3) For the output ref cursor that represents the return value for your Entity Function, choose "Add to Config" checkbox.
4) Then select either "Show Config" (and then cut and paste) or "Add to Config".
Here is a screenshot of what I am talking about:
http://i.imgur.com/t1BfmUP.gif
If this doesn't fix the problem, play around with that boolean mapping. I am not 100% sure of this as of this writing, but I remember hearing that support for booleans is another difference between managed and unmanaged ODP.NET. I'm sure it's buried in the release notes or doc somewhere.
Christian Shay
Oracle
Two things you would want to try which might potentially solve the issue:
Ensure the case of the schema name, stored procedure name and the
column names in the config are the same as that in the Oracle.
Try mapping the native type to a more conformant provider type, like
the first column COL1 - map an int32 providerType to the
number(10,0) nativeDataType as enforced by your edmmapping, instead of
the Decimal that you currently have. And so forth for the other
columns (like remove the column lengths) until you do not see the error or get a different one.
I've got the same error and I think my problem is a providerType of DOUBLE or DECIMAL. But, I got one to work that has your 3 column types. Your problem is that a number(10,0) should be a providerType of "Int64".
Stored Procedure:
create or replace PROCEDURE "PROC_ESCC_FIELDS" (p_recordset OUT SYS_REFCURSOR)
AS
BEGIN
OPEN p_recordset FOR
SELECT COL1, COL2, COL3
FROM MyTable;
END PROC_ESCC_FIELDS;
This works and returns the cursor:
<oracle.manageddataaccess.client>
<version number="*">
<implicitRefCursor>
<storedProcedure schema="SERFIS" name="PROC_V_SERFIS_ESCC_FIELDS">
<refCursor name="P_RECORDSET">
<bindInfo mode="Output" />
<metadata columnOrdinal="0" columnName="COL1" providerType="Int64" allowDBNull="false" nativeDataType="Number" />
<metadata columnOrdinal="1" columnName="COL2" providerType="Date" allowDBNull="true" nativeDataType="Date" />
<metadata columnOrdinal="2" columnName="COL3" providerType="Varchar2" allowDBNull="false" nativeDataType="Varchar2" />
</refCursor>
</storedProcedure>
</implicitRefCursor>
</version>
</oracle.manageddataaccess.client>
Click here for a list of the providerType and nativeDataType, etc. ENUMS:
I am just messing with Azure, and I can't seem to get my Db to work. I am following what it says here: https://www.windowsazure.com/en-us/develop/net/tutorials/web-site-with-sql-database/ and I updated my web.config to have this:
<entityFramework>
<defaultConnectionFactory type="System.Data.Entity.Infrastructure.SqlConnectionFactory, EntityFramework" />
<contexts>
<context type="DownloadThis.Models.DownloadThisDb, DownloadThisDb">
<databaseInitializer type="System.Data.Entity.MigrateDatabaseToLatestVersion">
<parameters>
<parameter value="DownloadThisDb_DatabasePublish" />
</parameters>
</databaseInitializer>
</context>
</contexts>
</entityFramework>
As is shown in the example, but I keep getting this error:
Format of the initialization string does not conform to specification
starting at index 0.
I have triple checked my connectionString, so that isn't it - any ideas?
Assuming that you're publishing with the VS2012 publish wizard, I've run into the same issue. If you choose to have the publishing wizard enable code first migrations instead of manually wiring them up in your code, then you need to provide a connection string in the publish settings. Your normal connection string is not used to run the migrations, this new connection string is used for this purpose only. This is nice because you can specify an account that has elevated privileges for performing your migrations, however, your app won't run under this user context. The problem is that the wizard doesn't make the need to specify this connection string very obvious. When you don't supply this, you end up with a null migration connection string and you spend a lot of time trying to figure out what's wrong with your normal connection strings.
Let's say that your context class is named FooContext. By convention, you'll have a connection string in your web.config named FooContext. When you enable code migrations via this wizard, the wizard will create a second connection string named FooContext_DatabasePublish that will only be used for running your code first migrations.
This blog post on MSDN explains this process in some detail.
I think you're missing a . in your type string:
<databaseInitializer type="System.Data.Entity.MigrateDatabaseToLatestVersion">
The red rectangle over the code makes it hard to read...
If that doest fix it, post a comment and I'll work up a sample to match yours and see if I can get it to work...
[UPDATED 2012-08-15]
OK - I think I know what's going on here... You mentioned "I updated my web.config to have this:" and showed your XML. When I ran through the tutorial, I did NOT have to enter ANY extra XML into my web.config. During the publishing process, the XML was added for me automagically by Visual Studio's deployment process and it all "just worked".
Here's your solution:
Go back to the original web.config file without these updates, and try publishing again.
For reference, here are the <entityFramework> sections from my two web.config files, first from my project, second from my hosted service (I got that by connecting to the running site via FTP and downloading it). The VS project is called 11964172 after the SO record number for this post:
local web.config file settings
<entityFramework>
<defaultConnectionFactory type="System.Data.Entity.Infrastructure.LocalDbConnectionFactory, EntityFramework">
<parameters>
<parameter value="v11.0" />
</parameters>
</defaultConnectionFactory>
</entityFramework>
deployed web.config file settings
<entityFramework>
<defaultConnectionFactory type="System.Data.Entity.Infrastructure.LocalDbConnectionFactory, EntityFramework">
<parameters>
<parameter value="v11.0" />
</parameters>
</defaultConnectionFactory>
<contexts>
<context type="_11963331.Models.ToDoDb, 11963331">
<databaseInitializer type="System.Data.Entity.MigrateDatabaseToLatestVersion`2[[_11963331.Models.ToDoDb, 11963331], [_11963331.Migrations.Configuration, 11963331]], EntityFramework, PublicKeyToken=b77a5c561934e089">
<parameters>
<parameter value="_11963331.Models.ToDoDb_DatabasePublish" />
</parameters>
</databaseInitializer>
</context>
</contexts>
</entityFramework>
I guess that explains why they took a picture of the web.config file changes instead of actually providing the code to type in :-)
See this question. If you leave the deployment connection string as "Remote connection string", and check the "Execute Code First Migrations" box, you will get this exception as the migration connection string, DownloadThisDb_DatabasePublish will not be defined properly. Either specify a real connection string instead of leaving the connection string box blank in the deployment wizard, or define a connection string named DownloadThisDb_DatabasePublish in your Azure site configuration.