I am trying to save objects in mongodb using multi tenancy at DATABASE level. However, objects always gets saved in the default database.
my application.yml
development:
grails:
mongodb:
tenantResolverClass: org.grails.datastore.mapping.multitenancy.resolvers.SystemPropertyTenantResolver
url: mongodb://localhost/test
connections:
1:
url: mongodb://localhost/test1db
2:
url: mongodb://localhost/test2db
options:
maxWaitTime: 10000
I can see that all the connections(default, 1 and 2) are loaded fine into connectionSourceMap in InMemoryConnectionSources.groovy class.
Interceptor function is invoked before controller function and interceptor sets the following property:
System.setProperty(SystemPropertyTenantResolver.PROPERTY_NAME, "1")
I was expecting that since Tenant id has been set to "1" and object against key "1" exists in connectionSourceMap, so now objects will be saved in test1db. However this is not the case when I try to save object
Tenants.withCurrent {
domainObj.save(flush: flush, validate: false)
}
I tried to debug the application.
MongoDataStore.java
#Override
public <T1> T1 withNewSession(Serializable tenantId, Closure<T1> callable) {
MongoDatastore mongoDatastore = getDatastoreForTenantId(tenantId);
Session session = mongoDatastore.connect();
try {
DatastoreUtils.bindNewSession(session);
return callable.call(session);
}
finally {
DatastoreUtils.unbindSession(session);
}
}
mongoDataStore object appears to contain the right information with default database string set to test1db. However, the session object is again referring to test db as default db. I could not understand why this is happening and how could I resolve this.
I have a feeling that issue might be related to the following point in the documentation[1]:
11.2.3. Multi Tenancy and the Session Factory
Note that if you reference the default SessionFactory or PlatformTransactionManager in your classes that are injected via Spring, these will not be tenant aware and will point directly to default data source.
But could not figure out on how to resolve the issue.
Thanks for your help and time. Any help is appreciated.
1: http://gorm.grails.org/latest/hibernate/manual/index.html#multiTenancy
Related
I try to add multi-tenancy support for my Quarkus app, following Quarkus hibernate-orm doc (see last section).
I have my CustomTenantResolver class and configure in application.properties, with multiple data sources, but no named persistent unit, see below:
# Default data source
quarkus.hibernate-orm.datasource=master
quarkus.hibernate-orm.database.generation=none
quarkus.hibernate-orm.multitenant=DATABASE
# ----- Tenant 'master' (default) ---------------
quarkus.datasource."master".db-kind=postgresql
quarkus.datasource."master".username=postgres
quarkus.datasource."master".password=password
quarkus.datasource."master".jdbc.url=jdbc:postgresql://localhost:5432/db_master
# ----- Tenant 'test' ---------------------------
quarkus.datasource.test.db-kind=postgresql
quarkus.datasource.test.username=postgres
quarkus.datasource.test.password=password
quarkus.datasource.test.jdbc.url=jdbc:postgresql://localhost:5432/db_test
Everything works fine for Web Services APIs functions - based on incoming web service calls, I can extract and supply tenant identifier for DB access.
Problem is, my app also needs to use callback method to listen on messages coming from Apache Pulsar queue. When a message comes in and triggers this callback, any DB access in this method will give this exception:
SessionFactory configured for multi-tenancy, but no tenant identifier specified: org.hibernate.HibernateException: SessionFactory configured for multi-tenancy, but no tenant identifier specified
at org.hibernate.internal.AbstractSharedSessionContract.<init>(AbstractSharedSessionContract.java:172)
at org.hibernate.internal.AbstractSessionImpl.<init>(AbstractSessionImpl.java:29)
at org.hibernate.internal.SessionImpl.<init>(SessionImpl.java:221)
at org.hibernate.internal.SessionFactoryImpl$SessionBuilderImpl.openSession(SessionFactoryImpl.java:1282)
at org.hibernate.internal.SessionFactoryImpl.openSession(SessionFactoryImpl.java:472)
at io.quarkus.hibernate.orm.runtime.session.TransactionScopedSession.acquireSession(TransactionScopedSession.java:86)
at io.quarkus.hibernate.orm.runtime.session.TransactionScopedSession.persist(TransactionScopedSession.java:138)
at io.quarkus.hibernate.orm.runtime.session.ForwardingSession.persist(ForwardingSession.java:53)
... (snipped)
Apparently my CustomTenantResolver class was not called during this listener callback as the callback is another fresh thread, hence no tenant id is supplied.
Do I miss anything? How about the scheduler in Quarkus - how does it support multi-tenancy in scheduled jobs?
Thanks for helps.
I had a similar issue when pulling messages from JMS. The cause of the issue is that io.quarkus.hibernate.orm.runtime.tenant.HibernateCurrentTenantIdentifierResolver ( which implements CurrentTenantIdentifierResolver and as the doc says Maps from the Quarkus {#link TenantResolver} to the Hibernate {#link CurrentTenantIdentifierResolver} model ) expects a request context to be active before calling our implementation of TenantResolver, as shown here:
// Make sure that we're in a request
if (!Arc.container().requestContext().isActive()) {
return null;
}
TenantResolver resolver = tenantResolver(persistenceUnitName);
String tenantId = resolver.resolveTenantId();
I solved it on my app by, first, enabling the request context on the JMS consumer:
Arc.container().requestContext().activate();
and, second, using a ThreadLocal to "pass" the current tenant id to the TenantResolver that will be called later by Hibernate ( through the HibernateCurrentTenantIdentifierResolver instance):
CurrentTenantLocal.setCurrentTenantId("public");
On my TenantResolver ( the class that implements TenantResolver ) I resolve the tenant from either an injected JsonWebToken jwt when it comes from a WebRequest, or using the ThreadLocal when consuming from JMS:
if ( CurrentTenantLocal.getCurrentTenantId() != null ) {
return CurrentTenantLocal.getCurrentTenantId();
}
Caveats:
Note that I haven't done an exhaustive search of the possible side effects of activating the request context... but I have no problems so far.
I'm trying to port some webjob code to the new Azure Functions. So far I've managed to import my DLL's and reference them succesfully, but when I use the connection string in my code, I get an error saying I have to add the ProviderName:
The connection string 'ConnectionString' in the application's
configuration file does not contain the required providerName
attribute."
Which is normally not a problem because in a webjob (or web app), this will be in the App or Web.config, and the connectionstring will simply be overwritten with whatever I entered in Azure.
With Azure Functions, I don't have a web.config (Although I tried adding one to no avail), so naturally the providername is missing.
How do I specify that?
EDIT:
After some playing around and some helpful tips by people below, I've almost managed to get it working.
What I do now is the following:
var connString = **MY CONN STRING FROM CONFIG**; // Constring without metadata etc.
EntityConnectionStringBuilder b = new EntityConnectionStringBuilder();
b.Metadata = "res://*/Entities.MyDB.csdl|res://*/Entities.MyDB.ssdl|res://*/Entities.MyDB.msl";
b.ProviderConnectionString = connString.ConnectionString;
b.Provider = "System.Data.SqlClient";
return new MyDB(b.ConnectionString);
Which gives me what I need for calling the database. I use a static method in a partial class to get an instance of the Database which runs the above code, and I decorate my MyDB Partial with [DbConfigurationType(typeof(MyDbConfiguration))]
I define that configuration as:
public class MyDBConfiguration: DbConfiguration
{
public MyDBConfiguration()
{
SetProviderFactory("System.Data.EntityClient", EntityProviderFactory.Instance);
}
}
My problem remains when I want to actually use the EF Entities. Here, it will try to initialize the database type using the original configuration, giving me the original error once again. As per this stack trace:
at Void Initialize()
at System.Data.Entity.Internal.EntitySetTypePair GetEntitySetAndBaseTypeForType(System.Type)
at Void InitializeContext()
at System.Data.Entity.Core.Objects.ObjectContext CreateObjectContextFromConnectionModel()
at Void Initialize()
at Boolean TryInitializeFromAppConfig(System.String, System.Data.Entity.Internal.AppConfig)
at Void InitializeFromConnectionStringSetting(System.Configuration.ConnectionStringSettings)
So how do I avoid this? I guess I need a way to hook into everything and run my custom setter..
In the end, Stephen Reindel pushed me in the right direction; Code-based Configuration for Entity Framework.
[DbConfigurationType(typeof(MyDBConfiguration))]
public partial class MyDB
{
public static MyDB GetDB()
{
var connString = **MY CONN STRING FROM SOMEWHERE**; // Constring without metadata etc.
EntityConnectionStringBuilder b = new EntityConnectionStringBuilder();
b.Metadata = "res://*/Entities.MyDB.csdl|res://*/Entities.MyDB.ssdl|res://*/Entities.MyDB.msl";
b.ProviderConnectionString = connString.ConnectionString;
b.Provider = "System.Data.SqlClient";
return new MyDB(b.ConnectionString);
}
public MyDB(string connectionString) : base(connectionString)
{
}
}
With MyDbConfiguration like this:
public class MyDBConfiguration: DbConfiguration
{
public MyDBConfiguration()
{
SetProviderServices("System.Data.SqlClient", SqlProviderServices.Instance);
SetDefaultConnectionFactory(new SqlConnectionFactory());
}
}
With the above code, EF never asks for AppConfig-related config files. But remember, if you have EF entries in your config file, it will attempt to use them, so make sure they're gone.
In terms of azure functions, this means I used the Azure Functions configuration panel in azure to punch in my ConnectionString without the Metadata and providername, and then loaded that in GetDB.
Edit: As per request, here is some explanatory text of the above:
You can't add EF metadata about the connection in Azure Functions, as they do not use an app.config in which to specify it. This is not a part of the connection string, but is metadata about the connection besides the connection string that EF uses to map to a specific C# Class and SQL Provider etc. To avoid this, you hardcode it using the above example. You do that by creating a class inheriting from DBConfiguration, and you mark (with an attribute on a partial class) your EF database class with that.
This DBConfiguration contains a different kind of way to instantiate a new database object, in which this metadata is hardcoded, but the connectionstring is retrieved from your app settings in Azure. In this example I just used a static method, but I guess it could be a new constructor also.
Once you have this static method in play, you can use that to get a new database in your database code, like this:
using (var db = MyDB.GetDB()) {
// db code here.
}
This allows you to use EntityFramework without an APP.Config, and you can still change the connectionstring using Azure Functions APP settings.
Hope that helps
Using this question you can set your default factory before opening the connection by having your personal DbConfiguration class (see this link also for usage):
public class MyDbConfiguration : DbConfiguration
{
public MyDbConfiguration()
{
SetDefaultConnectionFactory(new SqlConnectionFactory());
}
}
Now you need to tell your DbContext to use the new configuration. As using web.config or app.config is no option, you may use an attribute to add the configuration:
[DbConfigurationType(typeof(MyDbConfiguration))]
public class MyContextContext : DbContext
{
}
Now using a connection string on your DbContext will use the SQL provider by default.
Provided answer is perfect and it helped me a lot but it is not dynamic as I dont want to hardcode my connectionstring. if you are working the slots in azure functions. I was looking for a solution where I can use more than 1 connection strings. Here is my alternative approach step by step for anybody else struggling with this problem.
most important thing is that we understand local.settings.json file
IS NOT FOR AZURE. it is to run your app in the local as the name is
clearly saying. So solution is nothing to do with this file.
App.Config or Web.Config doesnt work for Azure function connection strings. If you have Database Layer Library you cant overwrite connection string using any of these as you would do in Asp.Net applications.
In order to work with, you need to define your connection string on the azure portal under the Application Settings in your Azure function. There is
Connection strings. there you should copy your connection string of your DBContext. if it is edmx, it will look like as below. There is Connection type, I use it SQlAzure but I tested with Custom(somebody claimed only works with custom) works with both.
metadata=res:///Models.myDB.csdl|res:///Models.myDB.ssdl|res://*/Models.myDB.msl;provider=System.Data.SqlClient;provider
connection string='data source=[yourdbURL];initial
catalog=myDB;persist security info=True;user
id=xxxx;password=xxx;MultipleActiveResultSets=True;App=EntityFramework
After you set this up, You need to read the url in your application and provide the DBContext. DbContext implements a constructor with connection string parameter. By default constructor is without any parameter but you can extend this. if you are using POCO class, you can amend DbContext class simply. If you use Database generated Edmx classes like me, you dont want to touch the auto generated edmx class instead of you want to create partial class in the same namespace and extend this class as below.
This is auto generated DbContext
namespace myApp.Data.Models
{
public partial class myDBEntities : DbContext
{
public myDBEntities()
: base("name=myDBEntities")
{
}
protected override void OnModelCreating(DbModelBuilder modelBuilder)
{
throw new UnintentionalCodeFirstException();
}
}
this is the new partial class, you create
namespace myApp.Data.Models
{
[DbConfigurationType(typeof(myDBContextConfig))]
partial class myDBEntities
{
public myDBEntities(string connectionString) : base(connectionString)
{
}
}
public class myDBContextConfig : DbConfiguration
{
public myDBContextConfig()
{
SetProviderServices("System.Data.EntityClient",
SqlProviderServices.Instance);
SetDefaultConnectionFactory(new SqlConnectionFactory());
}
}
}
After all you can get the connection string from azure settings, in your Azure Function project with the code below and provide to your DbContext
myDBEntities is the name you gave in the azure portal for your connection string.
var connString = ConfigurationManager.ConnectionStrings["myDBEntities"].ConnectionString;
using (var dbContext = new myDBEntities(connString))
{
//TODO:
}
Adding an answer in the event you cannot simply change the way you instantiate you DbContext. This would occur if you are calling code that has DbContexts being instatiated with the parameter-less constructor.
It involves using a static constructor to read your connection string from the appsettings in the azure portal and passing it in to your DbContext base constructor. This allows you to circumvent the need for a providerName and also allows you to retain use of the portal configuration without needing to hardcode anything.
Please refer to my accepted answer here: Missing ProviderName when debugging AzureFunction as well as deploying azure function
Stumbled upon this and solved it like this, inside of the Azure Function.
public static class MyFunction
{
// Putting this in more than one place in your project will cause an exception,
// if doing it after the DbConfiguration has been loaded.
static MyFunction() =>
DbConfiguration.Loaded += (_, d) =>
d.AddDefaultResolver(new global::MySql.Data.Entity.MySqlDependencyResolver());
// The rest of your function...
//[FunctionName("MyFunction")]
//public static async Task Run() {}
}
You can access the site's App Settings by going to the portal, clicking Function app settings and then Configure app settings. That will open up a blade that allows you to set all the app settings for your function app. Just use the same key and value that you'd use for your web.config.
I am writing a CLI app with Mybatis. In my app, when i go to different menus, it prompts for the user and password for the particular database that menu goes against.
I want to use Guice and Mybatis to handle all this but i have a slight problem. I want to use the XML config file to handle the mybatis config per database, but the user and pass from each connection has to come from the UI. So basically, i want to load mybatis xml file for a particular connection, then insert the credentials for the particular connection the user typed in, then bind those to the guice injector for that set of menus.
I can do it in java with a property object pretty easy, but i can't figure out how to do it with loading the XML first, then augmenting it with certain settings before loading.
Has anyone tried this?
If you are using mybatis guice this can be done by providing your dataSourceProvider for MyBatisModule like this:
Class<? extends Provider<DataSource>> dataSourceProviderType = YourDataSourceProvider.class;
Injector injector = Guice.createInjector(
new MyBatisModule() {
#Override
protected void initialize() {
bindDataSourceProviderType(dataSourceProviderType);
// other initialization ...
}
},
// other modules
);
YourDataSourceProvider should be able to create DataSource using credentials gotten from UI.
In this case you still can use xml mappers for mybatis.
So I'm messing around with SailsJS to try to get an API up and running real fast. I haven't setup a data store yet (will probably use mongodb) but I saw that there is what I'm assuming is like a SQLite database or something. So I generated a model and controller for a User. So in the browser I hit user/create. I see createdAt and updatedAt but no Id. Do I need a real datastore to see an ID? Is this still something you get for free?
Adapter
// Configure installed adapters
// If you define an attribute in your model definition,
// it will override anything from this global config.
module.exports.adapters = {
// If you leave the adapter config unspecified
// in a model definition, 'default' will be used.
'default': 'disk',
// In-memory adapter for DEVELOPMENT ONLY
// (data is NOT preserved when the server shuts down)
memory: {
module: 'sails-dirty',
inMemory: true
},
// Persistent adapter for DEVELOPMENT ONLY
// (data IS preserved when the server shuts down)
// PLEASE NOTE: disk adapter not compatible with node v0.10.0 currently
// because of limitations in node-dirty
// See https://github.com/felixge/node-dirty/issues/34
disk: {
module: 'sails-dirty',
filePath: './.tmp/dirty.db',
inMemory: false
},
// MySQL is the world's most popular relational database.
// Learn more: http://en.wikipedia.org/wiki/MySQL
mysql: {
module : 'sails-mysql',
host : 'YOUR_MYSQL_SERVER_HOSTNAME_OR_IP_ADDRESS',
user : 'YOUR_MYSQL_USER',
password : 'YOUR_MYSQL_PASSWORD',
database : 'YOUR_MYSQL_DB'
}
};
Model
/*---------------------
:: User
-> model
---------------------*/
module.exports = {
attributes: {
firstName: 'STRING',
lastName: 'STRING'
}
};
Do .../user/findAll to list all the users. Each user should have an 'id' property, e.g.
{
"username":"admin",
"password":"$2a$10$SP/hWtlJINkMgkAuAc9bxO1iWsvVcglwEU4ctGCiYpx.7ykaFWt56",
"createdAt":"2013-07-12T17:36:29.852Z",
"updatedAt":"2013-07-12T17:36:29.852Z",
"id":1
}
they removes findAll you can simply use find instad
u can also request like this http://example.de/Model/create?firstname=Foo&lastname=bar
I have checkout this by switching between different DB in a same sails application and its working absolutely fine.
So In this case you can add a property of autoPK to your model and setting it to true like following:
module.exports = {
attribute:{
//your model attributes with validations.
},
autoPK:true
}
If this doesn't works for you then I think there is something wrong with your copy of sails you installed on your system.
Please try to install the newer version (v0.9.8) of sails using npm or try to update your package.json with newer version (v0.9.8) details and excute npm install.
I have a web app that uses EF5 to map to a SQL database. This is the standard membership database with some additional tables I've added. Works like a champ in that project.
I have a second project, a windows service running TCP a server, which needs to insert items into the same database. So I reference the web app from this second project and can see my DbContext and entity types as needed.
At runtime, however, none of my DbSets gets populated with data. I have tried explicitly opening the connection to execute queries too, like this:
public MyContext()
: base("DefaultConnection")
{
try
{
Database.Connection.Open();
var command = new System.Data.SqlClient.SqlCommand("SELECT * FROM dbo.Trackers", (SqlConnection) Database.Connection);
var reader = command.ExecuteReader();
bool result = reader.Read();
}
catch (Exception exception)
{
//handle exception
}
this.Database.Connection.Close();
}
The result is false, but the connection is created and the reader is aware that I have four fields in my table. Is anyone aware of a reason this should work in my web app but not in a referencing app?
I had forgotten that the connection string from the web.config file is only honored when running as a web app. The TCP service .exe needs its own copy of the connection string in App.config. It just happened that the default (implicit) connection string on my TCP service connected to a valid, but empty, copy of my database.