I have app backend on php. But recently i created same backend on Quarkus and deployed it on kubernetes with database MYSQL. I'm routing half users on php backend and half qaurkus from app. PHP backend working fine but have issue with quarkus backend it stuck after few hours and request latency increase more than one minute and sometimes timeout. Well after checking some stacks mostly i see that it stuck at getConnection in datasource I'm suing agroal. Its not database issue because meanwhile if i check app with php that is working fine without any issue or delay. Also i tried increasing and decreasing min/max pool size but still same issue.
I will share my structure of using quarkus, agroal and jooq with mysql. I think issue is with that.
I'm using kotlin and DSLContext is object which is static.
Application.kt
#ApplicationScoped
class AppLifecycleBean {
#Inject
lateinit var dsl: DSLContext
#Inject
lateinit var launchMode: LaunchMode
fun onStart(#Observes ev: StartupEvent?) {
DBHelper.init(dsl)
}
fun onStop(#Observes ev: ShutdownEvent?) {
}
}
DBHelper.kt
object DBHelper {
lateinit var db: DSLContext
private set
fun init(db: DSLContext){
this.db = db
}
}
As you can see I'm initializing JOOQ at start of application as singleton and then setting its instance in object in DBHelper to keep using it instead creating new. So i doubt may be that is issue do i need to create new instance of DSLContet everytime by injecting. Then i created models to access database like this
open class Model {
protected val db
get() = DBHelper.db
}
object UserModel: Model() {
fun getUserbyId(id: Int){
db.selectFrom.......
}
}
At last I'm accessing models like this
#Path("/mobile/user")
#Produces(MediaType.APPLICATION_JSON)
class UserApiController() : Controller() {
#GET
#Path("/get_profile")
fun getProfile(
#QueryParam("user_id") userId: Int
): Response {
val user = UserModel.getbyId(userId)
return Response.success(data = user)
}
}
I think some where resources leak
Related
We currently have a very weird problem with a web application running inside Open Liberty 18.0.0.4. We are injecting a Mongo database connection into a JAX-RS request handler. The producer for the injected value simply returns a member variable that can never get null once its initialized. The service runs fine for many hours, sometimes even days and then suddenly any access to the injected Mongo database throws a NPE.
Here's a rough sketch of the code:
#ApplicationScoped
public class BackingServiceConnectionFactory {
private MongoDatabase m_mongoDatabase = ...;
#Produces
public MongoDatabase getMongoDatabase() {
return m_mongoDatabase;
}
}
#Path("repository")
public class RepositoryImpl {
#Inject
private MongoDatabase m_database;
#GET
public Response foo() {
MongoCollection<Document> workflows = m_database.getCollection("workflows"); // <== NPE
}
}
Once the injected variable is null all subsequent injections will also lead to null rendering the service unusable.
Any idea what the problem could be or how to debug it?
I am trying to create service that will read some data from remote server and process them using Spring Integration.
I have class that extends ArrayList, because I need to keep pointer to other page, so I can read it in next remote call. I set up release strategy to collect all these pages, until there is no pointer for the next page.
Here is definition of class:
public class CustomList extends ArrayList<DataInfo>
{
private String nextCursor;
// Methods omitted for readability
}
Everything worked fine, until I setup JdbcMessageStore in Aggregator, so I can keep messages in case of service shutdown.
Problem on which I come across is that in my release strategy class I cast my list class to same class (because message group does not define type), this exceptions is raised:
java.lang.ClassCastException: com.example.CustomList cannot be cast to com.example.CustomList
This is my release strategy class:
#Component
public class CursorReleaseStrategy implements ReleaseStrategy
{
#Override
public boolean canRelease(MessageGroup group)
{
return group.getMessages().stream()
.anyMatch(message -> ((CustomList) message.getPayload()).getNextCursor() == null);
}
}
If I remove message store, everything works fine, but the problem is that I need message store.
I am using spring boot 2.1.6 and Spring Integration DSL for creating this flow.
From what I read, this error happens because of different class loaders, but this I do from the same application.
Is there anything more that I need to configure for this to work_
Almost certainly a class loader issue; you can find which class loader loads each component (message store, release strategy) by injecting them into a bean and calling getClass().getClassLoader().
When application has been packaged in jar, there was such error.
So to fix the problem I created two beans, depending on profile.
For example:
#Profile("!prod")
#Bean
public MessageGroupStore messageStore(DataSource dataSource)
{
JdbcMessageStore jdbcMessageStore = new JdbcMessageStore(dataSource);
jdbcMessageStore.setDeserializer(inputStream -> {
ConfigurableObjectInputStream objectInputStream = new ConfigurableObjectInputStream(inputStream, Thread.currentThread().getContextClassLoader());
try {
return (Message<?>) objectInputStream.readObject();
} catch (ClassNotFoundException var4) {
throw new NestedIOException("Failed to deserialize object type", var4);
}
});
return jdbcMessageStore;
}
#Profile("prod")
#Bean
public MessageGroupStore prodMessageStore(DataSource dataSource)
{
return new JdbcMessageStore(dataSource);
}
ServiceStack.Funq.Quartz Sample Code is
public class MyServices : Service
{
public object Any(Hello request)
{
return new HelloResponse { Result = "Hello, {0}!".Fmt(request.Name) };
}
}
public class HelloJob : IJob
{
private MyServices MyServices { get; set; }
public HelloJob(MyServices myServices)
{
MyServices = myServices;
}
public void Execute(IJobExecutionContext context)
{
var response = MyServices.Any(new ServiceModel.Hello
{
Name = "CodeRevver"
});
response.PrintDump();
}
}
The above is works fine. if I in the MyServices Class, removed the Any function, and comment the Execute inner code.
public class MyServices : Service
{
}
the quartz.net will get the error:
[Quartz.Core.ErrorLogger】 An error occurred instantiating job to be executed. job= 'JobGroup1.GetUserJob111' Problem instantiating type 'ServiceStackWithQuartz.HelloJob'
why the class must have public object Any(Hello request) function ?
Thanks for using the package – I had no idea that other people would find it useful.
So If I understand correctly, in your situation you have:
public class MyServices : Service
{
}
And you’re trying to resolve this Service via constructor injection, which is effectively doing a:
container.Resolve<MyServices>();
This will fail because of the way the ServiceStack Funq IoC works. You can’t resolve a ServiceStack Service that has nothing in it (you'd probably never want to either) – It has to at least have one service implementation, It doesn’t matter what the implementation is.
Also, if you want to improve ServiceStack.Funq.Quartz, feel free to contribute to the code base.
Edit: It's probably worth mentioning that you can inject a "Non-Service" class with your logic in it if you want. You can resolve other classes that aren't based off of ServiceStack.Service even if there's nothing in them.
Edit 2: Responding to your "Service wont dispose" problem. This is the same across ServiceStack and has nothing to do with your Quartz Job. If you call a:
container.Resolve<MyServices>().Any(new new ServiceModel.Hello { });
from AppHost for example, your service wont dispose by itself. If you want it to dispose you can wrap it in a using statement. e.g.
using (var service = MyServices)
{
var response = MyServices.Any(new ServiceModel.Hello { });
}
The using will ensure that your service will be disposed afterwards.
Alternatively you can add the interface "IDispose" on to your Quartz Job and implement a Dispose() method that will do a:
MyServices.Dispose();
This will be called after a job has executed.
I am an Asp.net developer but very much new to the Asp.net Identity framework. I have been studying the sample application and followed some tutorials too on Identity but still I am not able to grasp the concept completely. I have very firm grip over Asp.net membership but Identity seems nothing like membership. I will explain what I have done so far.
I am creating a simple application in which I am following code first approach. I have created entity model for User which inherits from IdentityUser and has some extra fields. Below is entity model for User.
public class User : IdentityUser
{
public int? CompanyID { get; set; }
public bool? CanWork { get; set; }
public bool? CanSearch { get; set; }
public Company Company { get; set; }
}
Now in the examples people use the name ApplicationUser but for my own purpose I have used name User. Also there is a method in User or ApplicationUser model which is,
public async Task<ClaimsIdentity> GenerateUserIdentityAsync(UserManager<User> manager)
{
CookieAuthenticationOptions.AuthenticationType
var userIdentity = await manager.CreateIdentityAsync(this, DefaultAuthenticationTypes.ApplicationCookie);
// Add custom user claims here
return userIdentity;
}
I am unable to understand the purpose of this method. Also from an example I have used the following model for Role,
public class Role : IdentityRole
{
public Role()
{
}
public Role(string roleName, string description)
: base(roleName)
{
this.Description = description;
}
public string Description { get; set; }
}
I understand that an extra field is added but I am unable to understand the purpose of overloaded constructor.
The above mentioned confusions are secondary. My primary confusion is that I am familiar that when I create entity models I use DbSet and DbContext and when I call any entity framework method to access the database, the database is created/drop created whichever scheme I am following.
In Identity which method is responsible for creating the Identity tables in the database? I have a IdentityConfig file in which I declare ApplicationUserManager and ApplicationSignInManager. I have also a Startup file. Previously I had only one Startup file in the App_Start folder and when I run the application and tried to accessed any Identity methods it gave me error and was not creating database. I then made the class as partial and created another partial class with same name at the root and then the exception was gone and tables were created. So Startup class is responsible for creating Identity tables? There are extra columns created automatically in the AspNetUsers like PhoneNumber, PhoneNumberConfirmed, TwoFactorEnabled. I don't need these extra columns. Can I remove these? Can I change the names of the Identity tables that are created?
I know these are very basic questions and not one question at all but if I was unable to find some basic tutorial or example for beginners then it would be very beneficial. What I have found are describing those things which I don't need or making me confuse. I want to understand and have control how Identity should work in my application but till now it seems to me that neither I am grasping it completely and nor being able to make is adjustable to my needs. Its like tutorials and example are teaching me how to make sentences but I am unable to understand the alphabets. :(
First of all you have to define the model - as you're doing - implementing the right interfaces.
Let's say you want to create a user for your application:
public class MyUser : IdentityUser<string, MyUserLogin, MyUserRole, MyUserClaim>
{
public string CompanyName { get; set; }
}
As you can see I've implemented the IdentityUser interface (namespace Microsoft.AspNet.Identity.EntityFramework).
I've specified what type of identifier I want to use for my primary key (string) and included my custom objects to manges login, roles and claims.
Now we can defined the role object:
public class MyRole : IdentityRole<string, MyUserRole>
{
}
Again there's a type and the class I've defined for the management of users belonging to to a role.
public class MyUserRole : IdentityUserRole<string>
{
}
MyUserLogin is going to implement IdentityUserLogin<string>.
MyUserClaim is going to implement IdentityUserClaim<string>.
As you can see each interface need a type for the primary key.
The second step is to create the user store:
public class MyUserStore: UserStore<MyUser, MyRole, string, MyUserLogin, MyUserRole, MyUserClaim>
{
public MyUserStore(MyContext context)
: base(context)
{
}
}
Again we have defined what user, role, login etc etc we want to use.
We need UserStore cause our UserManager is going to need one.
If you're planning to manage roles and associate roles with each user you have to create your RoleStore definition.
public class MyRoleStore : RoleStore<MyRole, string, MyUserRole>
{
public DaufRoleStore(ApplicationDatabaseContext context) : base(context)
{
}
}
Now you can create your UserManager. The UserManager is the real responsible of saving changes to the UserStore.
public class ApplicationUserManager : UserManager<MyUser, string>
{
public ApplicationUserManager(IUserStore<MyUser, string> store)
: base(store)
{
}
public static ApplicationUserManager Create(IdentityFactoryOptions<ApplicationUserManager> options, IOwinContext context)
{
var manager = new ApplicationUserManager(new MyUserStore(context.Get<MyContext>()));
manager.UserValidator = new UserValidator<MyUser, string>(manager)
{
AllowOnlyAlphanumericUserNames = false,
RequireUniqueEmail = true
};
manager.PasswordValidator = new PasswordValidator()
{
RequiredLength = 5,
RequireNonLetterOrDigit = false, // true
// RequireDigit = true,
RequireLowercase = false,
RequireUppercase = false,
};
return (manager);
}
}
This class has a static method which will create a new UserManager for you.
Interesting to note that you can include some validation rules you might need to validate password etc etc.
Last thing is to create or database context.
public class MyContext : IdentityDbContext<MyUser, MyRole, string, MyUserLogin, MyUserRole, MyUserClaim>
{
public MyContext(): base("<your connection string here>")
{
}
public static MyContext Create()
{
return new MyContext();
}
protected override void OnModelCreating(DbModelBuilder modelBuilder)
{
base.OnModelCreating(modelBuilder);
modelBuilder.Entity<MyUser>()
.ToTable("Users");
modelBuilder.Entity<MyRole>()
.ToTable("Roles");
modelBuilder.Entity<MyUserRole>()
.ToTable("UserRoles");
modelBuilder.Entity<MyUserClaim>()
.ToTable("UserClaims");
modelBuilder.Entity<MyUserLogin>()
.ToTable("UserLogins");
}
}
As you can see I've used the model builder to change the names all the tables.
You can define keys or fields type or tables relations here.
This is the place where you're going to attach your custom classes you want to manage in your context:
public DbSet<MyCustomer> Customers{ get; set; }
Again MyContext has a Create method which returns a new context:
public static MyContext Create()
{
return new MyContext();
}
Now you should have a startup class where you're going to bootstrap your stuff:
[assembly: OwinStartup(typeof(ASPNETIdentity2.Startup))]
namespace ASPNETIdentity2
{
public class Startup
{
public void Configuration(IAppBuilder app)
{
app.CreatePerOwinContext(MyContext.Create);
app.CreatePerOwinContext<ApplicationUserManager>(ApplicationUserManager.Create);
}
}
}
Here you're going to create your database context and your user manager you can use in your application.
Notice the first line:
[assembly: OwinStartup(typeof(ASPNETIdentity2.Startup))]
This is needed cause you're telling your environment that is the startup class which needs to be called at ... startup.
Now in your controllers you can simply refer to your UserManager doing something like this:
HttpContext.GetOwinContext().GetUserManager<ApplicationUserManager>();
How can you create your tables?
In Visual Studio go to TOOLS -> NuGet Packager Manager -> Package Manager Console.
In the window there's a combobox "Default Project". Choose your ASP.NET MVC project.
Run this command:
Enable-Migrations
It will create a file Configuration.cs in a new folder called Migrations.
If you want to create your database you need to open that file and change the AutomaticMigrationsEnabled to true:
public Configuration()
{
AutomaticMigrationsEnabled = true;
}
Again, from Package Manager Console, you can run:
Update-Database
and all your tables will appear in your database. Don't forget your connection string.
You can download this github project to see how everything works.
You can check these two answers with some other info.
The first of the two has got some links to a blog where you can learn all these things.
NOTE:
You have to do all this if you want to customized every single bit of your environment.
I'm new to SubSonic (of all flavours), but thought I might as well start with 3.0, because I'd like to use Linq, and I get the impression 3.0 is not that far away from release.
I tried the alpha download .zip, but that seems pretty old and didn't singularize table class names, so I'm now running from the latest trunk SVN version (rev62).
I've run the 'simple' templates, from SubSonic.Templates\Simple against my database and everything seems ok, but the DB context class which the templates create starts like this:
public partial class DB : IQuerySurface
{
static DB _db;
public DB() {
_db = new DB();
}
public static DB CreateDB()
{
if (_db == null)
{
_db = new DB();
_db.Init();
}
return _db;
}
... etc
Unsurprisingly, when I call DB.CreateDB, the ctor recurses endlessly and crashes everything with a stack overflow.
I don't really understand the ctor at all - it doesn't look like it should be there, but both the 'simple' and the 'advanced' templates create something similar, and there's a humongous test suite which I imagine is verifying this somehow.
Clearly I have the wrong end of the stick here - what blindingly obvious point have I missed?
Update: The simple and advanced templates aren't similar, and the advanced ones don't have this problem. Thanks for the help.
Another Update: It looks like this is fixed in the simple templates in SVN r66
Don't know if you have the latest bits from SVN with a bug, but my version from a few days ago appears to be working fine. Here is what my DB class starts off with:
public partial class DB : IQuerySurface
{
BatchQuery _batch = null;
public IDataProvider DataProvider;
public DbQueryProvider provider;
private IDatabaseSchema _schema;
public IDatabaseSchema Schema
{
get
{
return _schema;
}
}
public DB()
{
DataProvider = ProviderFactory.GetProvider("Northwind");
Init();
}
public DB(string instanceName, string connectStr)
{
SubSonic.DataProviders.ConnectionStringProvider.Instance.AddLocalConnectionString(
instanceName, connectStr, "System.Data.SqlClient");
DataProvider = ProviderFactory.GetProvider(instanceName);
Init();
}
... etc...
I used the advanced version of the templates.
I prefer the t4 templates, here is the ctor provided there:
public DB()
{
DataProvider = ProviderFactory.GetProvider("Northwind");
Init();
}
there is also an overload that accepts a connection string. This is working quite well for me, I'm using the linq support and it is full of awesome.