Second level caching is a query cache. The results of EF commands will be stored in the cache, so that the same EF commands will retrieve their data from the cache rather than executing them against the database again.
To install EFCoreSecondLevelCacheInterceptor, run the following command in the Package Manager Console:
PM> Install-Package EFCoreSecondLevelCacheInterceptor
You can also view the package page on NuGet.
You can use the following cache providers:
- Memory provider built in this library
- EasyCaching.Core memory provider
- EasyCaching.Core dynamic cache provider
- CacheManager.Core cache provider
- A custom cache provider
namespace EFCoreSecondLevelCacheInterceptor.AspNetCoreSample
{
public class Startup
{
private readonly string _contentRootPath;
public Startup(IConfiguration configuration, IWebHostEnvironment env)
{
_contentRootPath = env.ContentRootPath;
Configuration = configuration;
}
public IConfiguration Configuration { get; }
public void ConfigureServices(IServiceCollection services)
{
services.AddEFSecondLevelCache(options =>
options.UseMemoryCacheProvider().DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
// Please use the `CacheManager.Core` or `EasyCaching.Redis` for the Redis cache provider.
);
var connectionString = Configuration["ConnectionStrings:ApplicationDbContextConnection"];
if (connectionString.Contains("%CONTENTROOTPATH%"))
{
connectionString = connectionString.Replace("%CONTENTROOTPATH%", _contentRootPath);
}
services.AddConfiguredMsSqlDbContext(connectionString);
services.AddControllersWithViews();
}
}
}
Here you can use the EasyCaching.Core, as a highly configurable cache manager too.
To use its in-memory caching mechanism, add this entry to the .csproj
file:
<ItemGroup>
<PackageReference Include="EasyCaching.InMemory" Version="1.6.1" />
</ItemGroup>
Then register its required services:
namespace EFSecondLevelCache.Core.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
const string providerName1 = "InMemory1";
services.AddEFSecondLevelCache(options =>
options.UseEasyCachingCoreProvider(providerName1, isHybridCache: false).DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
);
// Add an in-memory cache service provider
// More info: https://easycaching.readthedocs.io/en/latest/In-Memory/
services.AddEasyCaching(options =>
{
// use memory cache with your own configuration
options.UseInMemory(config =>
{
config.DBConfig = new InMemoryCachingOptions
{
// scan time, default value is 60s
ExpirationScanFrequency = 60,
// total count of cache items, default value is 10000
SizeLimit = 100,
// enable deep clone when reading object from cache or not, default value is true.
EnableReadDeepClone = false,
// enable deep clone when writing object to cache or not, default value is false.
EnableWriteDeepClone = false,
};
// the max random second will be added to cache's expiration, default value is 120
config.MaxRdSecond = 120;
// whether enable logging, default is false
config.EnableLogging = false;
// mutex key's alive time(ms), default is 5000
config.LockMs = 5000;
// when mutex key alive, it will sleep some time, default is 300
config.SleepMs = 300;
}, providerName1);
});
}
}
}
If you want to use the Redis as the preferred cache provider with EasyCaching.Core
, first install the following package:
<ItemGroup>
<PackageReference Include="EasyCaching.Redis" Version="1.6.1" />
<PackageReference Include="EasyCaching.Serialization.MessagePack" Version="1.6.1" />
</ItemGroup>
And then register its required services:
namespace EFSecondLevelCache.Core.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
const string providerName1 = "Redis1";
services.AddEFSecondLevelCache(options =>
options.UseEasyCachingCoreProvider(providerName1, isHybridCache: false).DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails (for example, if Redis is down).
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
);
// More info: https://easycaching.readthedocs.io/en/latest/Redis/
services.AddEasyCaching(option =>
{
option.UseRedis(config =>
{
config.DBConfig.AllowAdmin = true;
config.DBConfig.SyncTimeout = 10000;
config.DBConfig.AsyncTimeout = 10000;
config.DBConfig.Endpoints.Add(new EasyCaching.Core.Configurations.ServerEndPoint("127.0.0.1", 6379));
config.EnableLogging = true;
config.SerializerName = "Pack";
config.DBConfig.ConnectionTimeout = 10000;
}, providerName1)
.WithMessagePack(so =>
{
so.EnableCustomResolver = true;
so.CustomResolvers = CompositeResolver.Create(
new IMessagePackFormatter[]
{
DBNullFormatter.Instance, // This is necessary for the null values
},
new IFormatterResolver[]
{
NativeDateTimeResolver.Instance,
ContractlessStandardResolver.Instance,
StandardResolverAllowPrivate.Instance,
});
},
"Pack");
});
}
}
}
If you want to support multitenancy in your application and have a different Redis database per each tenant, first register multiple pre-configured providers with known providerName
s and then select these providerName
s based on the current tenant this way dynamically:
services.AddEFSecondLevelCache(options =>
options.UseEasyCachingCoreProvider(
(serviceProvider, cacheKey) => "redis-db-" + serviceProvider.GetRequiredService<IHttpContextAccesor>().HttpContext.Request.Headers["tenant-id"],
isHybridCache: false)
// `Or` you can set the cache key prefix per tenant dynamically
.UseCacheKeyPrefix(serviceProvider => "EF_" + serviceProvider.GetRequiredService<IHttpContextAccesor>().HttpContext.Request.Headers["tenant-id"])
.DisableLogging(true)
.UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
);
Also here you can use the CacheManager.Core, as a highly configurable cache manager too.
To use its in-memory caching mechanism, add these entries to the .csproj
file:
<ItemGroup>
<PackageReference Include="CacheManager.Core" Version="1.2.0" />
<PackageReference Include="CacheManager.Microsoft.Extensions.Caching.Memory" Version="1.2.0" />
<PackageReference Include="CacheManager.Serialization.Json" Version="1.2.0" />
</ItemGroup>
Then register its required services:
namespace EFSecondLevelCache.Core.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
services.AddEFSecondLevelCache(options =>
options.UseCacheManagerCoreProvider().DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
);
// Add an in-memory cache service provider
services.AddSingleton(typeof(ICacheManager<>), typeof(BaseCacheManager<>));
services.AddSingleton(typeof(ICacheManagerConfiguration),
new CacheManager.Core.ConfigurationBuilder()
.WithJsonSerializer()
.WithMicrosoftMemoryCacheHandle(instanceName: "MemoryCache1")
.Build());
}
}
}
If you want to use the Redis as the preferred cache provider with CacheManager.Core
, first install the CacheManager.StackExchange.Redis
package and then register its required services:
// Add Redis cache service provider
var jss = new JsonSerializerSettings
{
NullValueHandling = NullValueHandling.Ignore,
ReferenceLoopHandling = ReferenceLoopHandling.Ignore,
TypeNameHandling = TypeNameHandling.Auto,
Converters = { new SpecialTypesConverter() }
};
const string redisConfigurationKey = "redis";
services.AddSingleton(typeof(ICacheManagerConfiguration),
new CacheManager.Core.ConfigurationBuilder()
.WithJsonSerializer(serializationSettings: jss, deserializationSettings: jss)
.WithUpdateMode(CacheUpdateMode.Up)
.WithRedisConfiguration(redisConfigurationKey, config =>
{
config.WithAllowAdmin()
.WithDatabase(0)
.WithEndpoint("localhost", 6379)
// Enables keyspace notifications to react on eviction/expiration of items.
// Make sure that all servers are configured correctly and 'notify-keyspace-events' is at least set to 'Exe', otherwise CacheManager will not retrieve any events.
// You can try 'Egx' or 'eA' value for the `notify-keyspace-events` too.
// See https://redis.io/topics/notifications#configuration for configuration details.
.EnableKeyspaceEvents();
})
.WithMaxRetries(100)
.WithRetryTimeout(50)
.WithRedisCacheHandle(redisConfigurationKey)
.Build());
services.AddSingleton(typeof(ICacheManager<>), typeof(BaseCacheManager<>));
services.AddEFSecondLevelCache(options =>
options.UseCacheManagerCoreProvider().DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
);
Here is the definition of the SpecialTypesConverter.
If you don't want to use the above cache providers, implement your custom IEFCacheServiceProvider
and then introduce it using the options.UseCustomCacheProvider<T>()
method.
2- Add SecondLevelCacheInterceptor to your DbContextOptionsBuilder
pipeline:
public static class MsSqlServiceCollectionExtensions
{
public static IServiceCollection AddConfiguredMsSqlDbContext(this IServiceCollection services, string connectionString)
{
services.AddDbContextPool<ApplicationDbContext>((serviceProvider, optionsBuilder) =>
optionsBuilder
.UseSqlServer(
connectionString,
sqlServerOptionsBuilder =>
{
sqlServerOptionsBuilder
.CommandTimeout((int)TimeSpan.FromMinutes(3).TotalSeconds)
.EnableRetryOnFailure()
.MigrationsAssembly(typeof(MsSqlServiceCollectionExtensions).Assembly.FullName);
})
.AddInterceptors(serviceProvider.GetRequiredService<SecondLevelCacheInterceptor>()));
return services;
}
}
Note: Some database providers don't support special fields such as DateTimeOffset
, TimeSpan
, etc. For these scenarios you will need the related converters.
This library doesn't need any settings for the cache invalidation. It watches for all of the CRUD operations using its interceptor and then invalidates the related cache entries automatically.
But if you want to invalidate the whole cache manually
, inject the IEFCacheServiceProvider
service and then call its _cacheServiceProvider.ClearAllCachedEntries()
method or use it this way to specify the root cache keys which are a collection of a Prefix+TableName:
// Partial cache invalidation using the specified table names
// This is useful when you are monitoring your DB's changes using the SqlTableDependency
_cacheServiceProvider.InvalidateCacheDependencies(new EFCacheKey(new HashSet<string>()
{
"EF_TableName1", // "EF_" is the cache key's prefix
"EF_TableName2"
} { KeyHash = "empty" }));
var post1 = context.Posts
.Where(x => x.Id > 0)
.OrderBy(x => x.Id)
.FirstOrDefault();
We can use the new Cacheable()
extension method:
var post1 = context.Posts
.Where(x => x.Id > 0)
.OrderBy(x => x.Id)
.Cacheable(CacheExpirationMode.Sliding, TimeSpan.FromMinutes(5))
.FirstOrDefault(); // Async methods are supported too.
NOTE: It doesn't matter where the Cacheable
method is located in this expression tree. It just adds the standard TagWith
method to mark this query as Cacheable
. Later SecondLevelCacheInterceptor
will use this tag to identify the Cacheable
queries.
Also it's possible to set the Cacheable()
method's settings globally:
services.AddEFSecondLevelCache(options => options.UseMemoryCacheProvider(CacheExpirationMode.Absolute, TimeSpan.FromMinutes(5)).DisableLogging(true)
.UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
);
In this case the above query will become:
var post1 = context.Posts
.Where(x => x.Id > 0)
.OrderBy(x => x.Id)
.Cacheable()
.FirstOrDefault(); // Async methods are supported too.
If you specify the settings of the Cacheable()
method explicitly such as Cacheable(CacheExpirationMode.Sliding, TimeSpan.FromMinutes(5))
, its setting will override the global setting.
To cache all of the system's queries, just set the CacheAllQueries()
method:
namespace EFCoreSecondLevelCacheInterceptor.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
services.AddEFSecondLevelCache(options =>
{
options.UseMemoryCacheProvider().DisableLogging(true).UseCacheKeyPrefix("EF_");
options.CacheAllQueries(CacheExpirationMode.Absolute, TimeSpan.FromMinutes(30));
// Fallback on db if the caching provider fails.
options.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1));
});
// ...
This will put the whole system's queries in cache. In this case calling the Cacheable()
methods won't be necessary. If you specify the Cacheable()
method, its setting will override this global setting. If you want to exclude some of the queries from this global cache, apply the NotCacheable()
method to them.
To cache some of the system's queries based on their entity-types or table-names, use CacheQueriesContainingTypes
or CacheQueriesContainingTableNames
methods:
namespace EFCoreSecondLevelCacheInterceptor.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
services.AddEFSecondLevelCache(options =>
{
options.UseMemoryCacheProvider().DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
/*.CacheQueriesContainingTypes(
CacheExpirationMode.Absolute, TimeSpan.FromMinutes(30), TableTypeComparison.Contains,
typeof(Post), typeof(Product), typeof(User)
)*/
.CacheQueriesContainingTableNames(
CacheExpirationMode.Absolute, TimeSpan.FromMinutes(30), TableNameComparison.ContainsOnly,
"posts", "products", "users"
);
});
// ...
This will put the the specified system's queries in cache. In this case calling the Cacheable()
methods won't be necessary. If you specify the Cacheable()
method, its setting will override this global setting. If you want to exclude some of the queries from this global cache, apply the NotCacheable()
method to them.
To skip caching some of the system's queries based on their SQL commands, set the SkipCachingCommands
predicate:
namespace EFCoreSecondLevelCacheInterceptor.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
services.AddEFSecondLevelCache(options =>
{
options.UseMemoryCacheProvider().DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
// How to skip caching specific commands
.SkipCachingCommands(commandText =>
commandText.Contains("NEWID()", StringComparison.InvariantCultureIgnoreCase));
});
// ...
To skip caching some of the system's queries based on their results, set the SkipCachingResults
predicate:
namespace EFCoreSecondLevelCacheInterceptor.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
services.AddEFSecondLevelCache(options =>
{
options.UseMemoryCacheProvider().DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
// Don't cache null values. Remove this optional setting if it's not necessary.
.SkipCachingResults(result =>
result.Value == null || (result.Value is EFTableRows rows && rows.RowsCount == 0));
});
// ...
To do not cache some of the system's queries based on their entity-types or table-names, use CacheAllQueriesExceptContainingTypes
or CacheAllQueriesExceptContainingTableNames
methods:
namespace EFCoreSecondLevelCacheInterceptor.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
services.AddEFSecondLevelCache(options =>
{
options.UseMemoryCacheProvider().DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
/*.CacheAllQueriesExceptContainingTypes(
CacheExpirationMode.Absolute, TimeSpan.FromMinutes(30),
typeof(Post), typeof(Product), typeof(User)
)*/
.CacheAllQueriesExceptContainingTableNames(
CacheExpirationMode.Absolute, TimeSpan.FromMinutes(30),
"posts", "products", "users"
);
});
// ...
This will not put the the specified system's queries in cache. In this case calling the Cacheable()
methods won't be necessary. If you specify the Cacheable()
method, its setting will override this global setting.
Sometimes you don't want to invalidate the cache immediately, such when you are updating a post's likes or views count. In this case to skip invalidating the related cache entries of a given CRUD command, set the SkipCacheInvalidationCommands
predicate:
namespace EFCoreSecondLevelCacheInterceptor.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
services.AddEFSecondLevelCache(options =>
{
options.UseMemoryCacheProvider().DisableLogging(true).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
.SkipCacheInvalidationCommands(commandText =>
// How to skip invalidating the related cache entries of this query
commandText.Contains("NEWID()", StringComparison.InvariantCultureIgnoreCase));
});
// ...
This library uses the XxHash64Unsafe class to calculate the hash of a query and its parameters to produce a corresponding cache-key.
xxHash
is an extremely fast non-cryptographic
Hash algorithm. If you don't like it or you want to change it, just implement the IEFHashProvider
interface and then introduce it this way:
namespace EFCoreSecondLevelCacheInterceptor.AspNetCoreSample
{
public class Startup
{
public void ConfigureServices(IServiceCollection services)
{
services.AddEFSecondLevelCache(options =>
{
options.UseCustomHashProvider<MyCustomHashProvider>();
// ...
});
// ...
You should enable the logging system to see the behind the scene of the caching interceptor.
First set the DisableLogging(false)
:
services.AddEFSecondLevelCache(options =>
options.UseMemoryCacheProvider().DisableLogging(false).UseCacheKeyPrefix("EF_")
// Fallback on db if the caching provider fails.
.UseDbCallsIfCachingProviderIsDown(TimeSpan.FromMinutes(1))
And then change the log level to Debug
in your appsettings.json
file:
{
"Logging": {
"LogLevel": {
"Default": "Debug",
"System": "Debug",
"Microsoft": "Debug",
"Microsoft.Hosting.Lifetime": "Debug"
}
}
}
Now after running a query multiple times, you should have these logged lines:
Suppressed result with a TableRows[ee20d2d7-ffc7-4ff9-9484-e8d4eecde53e] from the cache[KeyHash: EB153BD4, CacheDependencies: Page.].
Using the TableRows[ee20d2d7-ffc7-4ff9-9484-e8d4eecde53e] from the cache.
Notes:
- Having the
Suppressed the result with the TableRows
message means the caching interceptor is working fine. - The next
Executed DbCommand
means nothing and it always will be logged by EF. - At the beginning there will be a lot of internal commands executed by the EF to run migrations, etc. Ignore these commands, because you will see the
Suppressed the result with the TableRows
messages for the frequently running queries. - Also you should verify it with a real DB profiler. It will log the 1st executed query and then on the 2nd run, you won't see it anymore.
Good candidates for query caching are global site settings and public data, such as infrequently changing articles or comments. It can also be beneficial to cache data specific to a user so long as the cache expires frequently enough relative to the size of the user base that memory consumption remains acceptable. Small, per-user data that frequently exceeds the cache's lifetime, such as a user's photo path, is better held in user claims, which are stored in cookies, than in this cache.
This cache is scoped to the application, not the current user. It does not use session variables. Accordingly, when retrieving cached per-user data, be sure queries in include code such as .Where(x => .... && x.UserId == id)
.
This cache is updated when an entity is changed (insert, update, or delete) via a DbContext that uses this library. If the database is updated through some other means, such as a stored procedure or trigger, the cache becomes stale.
To avoid complications, all of the queries inside an explicit
transaction (context.Database.BeginTransaction()) will not be cached. But the cache invalidations due to its CRUD operations will occur. You can use .AllowCachingWithExplicitTransactions(true)
setting to disable it.