我有一个在多个 WebAPI 服务上运行 ApacheIgnite 2.7.6 的设置。我在 Global.asax Application_Start 方法中为每个 WebAPI 服务实例化 IIGnite。然后在我的控制器中，我从WebApplication 类中的公共DataCache属性访问缓存。服务相互调用，Ignite 为所有服务创建一个公共缓存。现在，当连续运行单元测试时，Ignite 变得不稳定。时间是随机的，有时30分钟，有时20小时。寻找使我的代码稳定的提示或关于如何在不稳定后恢复的提示。
我的设置代码包括一个用于实例化 IIgnite 的循环，看起来像这样：
using Apache.Ignite.Core;
using Apache.Ignite.Core.Cache;
using Apache.Ignite.Core.Resource;
using Apache.Ignite.Linq;
namespace MyService
{
public class WebApiApplication : System.Web.HttpApplication
{
[InstanceResource]
private static IIgnite IgniteEngine; //{ get; set; } = null;
private static Apache.Ignite.Core.Cache.Configuration.QueryEntity queryEntity = new Apache.Ignite.Core.Cache.Configuration.QueryEntity( typeof( Tuple<string, byte> ), typeof( SomeClass ) );
public static ICache<Tuple<string, byte>, SomeClass> DataCache => IgniteEngine.GetOrCreateCache<Tuple<string, byte>, SomeClass>( new Apache.Ignite.Core.Cache.Configuration.CacheConfiguration( nameof( DataCache ), queryEntity ) );
protected void Application_Start( object sender, EventArgs e )
{
GlobalConfiguration.Configure(WebApiConfig.Register);
InitializeCache();
}
protected void Application_End( object sender, EventArgs e ){}
public static void InitializeCache()
{
try
{
short startCounter = 0;
Stopwatch sw = new Stopwatch();
sw.Start();
//Get the Ignite name - give options for debug environment and production environment
string siteName = System.Web.Hosting.HostingEnvironment.SiteName ?? System.Reflection.Assembly.GetExecutingAssembly()?.GetName()?.Name ?? System.AppDomain.CurrentDomain?.FriendlyName;
#if ( DEBUG || UNIT_TEST )
//Ignite uses H2 Database internally. Cache data is represented as SQL tables based on the CacheConfiguration
//To open H2 debug console and examine the database at runtime, set the IGNITE_H2_DEBUG_CONSOLE environment variable to true before starting Ignite.
//A web browser window will open during Ignite startup and show the state of the H2 database on the current node. Wait for the caches to start and refresh the browser.
Environment.SetEnvironmentVariable( "IGNITE_H2_DEBUG_CONSOLE", "true" );
#endif
//The JVM takes time to start. It could take several minutes when debugging
do
{
startCounter++;
//Sleep a little between tries
#if ( DEBUG || UNIT_TEST )
if ( startCounter >= 1 ) System.Threading.Thread.Sleep( 2000 );
#endif
try
{//Try to get a pre-existing instance of Ignite
IgniteEngine = Ignition.TryGetIgnite( siteName );
if ( null == IgniteEngine )
{
Debug.WriteLine( "InitializeCache attempt {0} to get an existing ignite instance failed.".Format( startCounter ) );
}
else
{
Debug.WriteLine( "InitializeCache attempt {0} to get an existing ignite instance succeeded.".Format( startCounter ) );
}
}
catch ( Exception ex )
{
Debug.WriteLine( "InitializeCache error while trying to get existing ignite instance. {0}".Format( ex.Message ) );
}
if ( null == IgniteEngine )
{
try
{
//Start the engine
string path = System.Web.HttpContext.Current?.Server?.MapPath( @"~\bin\libs" ) ?? System.IO.Path.GetDirectoryName( System.Reflection.Assembly.GetExecutingAssembly().Location ) + @"\libs";
string workPath = @"C:\Dev\Somefolder";
IgniteEngine = Ignition.Start( new IgniteConfiguration
{
JvmClasspath = System.IO.Directory.GetFiles( path ).Aggregate( ( x, y ) => x + ";" + y ),
IgniteInstanceName = siteName,
//Logger = new LogHelperForApacheIgnite(),
WorkDirectory = workPath
} );
if ( null != IgniteEngine )
{
Debug.WriteLine( "InitializeCache success starting Apache Ignite after {0} attempts and {1} seconds".Format( startCounter, sw.Elapsed.TotalSeconds ) );
}
}
catch ( System.BadImageFormatException ex1 )
{
Debug.WriteLine( "InitializeCache failed while trying to start a new ignite instance. The Apache Ignite cache runs as a 64bit dll but an attempt was made to invoke it while running the unit test at 32bits. Fix this from the Test menu by selecting \"Test Settings -> Default Processor Architecture -> x64\". The full error message is: {0}".Format( ex1.Message() ) );
}
catch ( Exception ex2 )
{
if (ex2.InnerException != null && ex2.InnerException.Message.Contains( "ERROR_BAD_EXE_FORMAT" ))
{
Debug.WriteLine( "InitializeCache failed while trying to start a new ignite instance. The Apache Ignite cache runs as a 64bit dll but an attempt was made to invoke it while running the unit test at 32bits. Fix this from the Test menu by selecting \"Test Settings -> Default Processor Architecture -> x64\". The full error message is: {0}".Format( ex2.InnerException?.Message ) );
}
else
{
Debug.WriteLine( "InitializeCache error while trying to start a new Apache Ignite instance. {0}".Format( ex2.Message ) );
}
}
}
}
while ( null == IgniteEngine && sw.Elapsed.TotalMinutes <= 4 );//The JVM sometimes takes more than a minute
sw.Stop();
if ( IgniteEngine == null )
{
string message = "Apache Ignite failed to start. InitializeCache unable to start ignite after {0} tries and {1} seconds".Format( startCounter, sw.Elapsed.TotalSeconds );
Debug.WriteLine( message );
throw new Exception( message );
}
else
{
Debug.WriteLine( "InitializeCache: Ignite running fine. Proceeding to create cached structure(s)." );
}
}
catch ( Exception ex )
{
Debug.WriteLine( "InitializeCache: Exception thrown during Cache initialization {0}".Format( ex.Message ) );
}
}
}
}
一旦 Ignite 变得不稳定，我已经看到了两种类型的错误：
(1) 当调用DataCache.ContainsKey( new Tuple( somestring, somebyte ) )我得到：
```BinaryObjectException "Failed to serialize object [typeName=String]"
Failed to serialize object [typeName=String]
EXCEPTION
Failed to serialize object [typeName=String]
at Apache.Ignite.Core.Impl.PlatformJniTarget.InStreamOutLong[TR](Int32 type, Action`1 outAction, Func`3 inAction, Func`2 readErrorAction)
at Apache.Ignite.Core.Impl.PlatformTargetAdapter.DoOutInOpX(Int32 type, Action`1 outAction, Func`2 inErrorAction)
INNER EXCEPTION
{"class org.apache.ignite.binary.BinaryObjectException: Failed to serialize object [typeName=String]\r\n\tat org.apache.ignite.internal.binary.BinaryClassDescriptor.write(BinaryClassDescriptor.java:840)\r\n\```
(2) 当调用DataCache.AsCacheQueryable().Where( x => x.Name == "Something" ).SingleOrDefault()或我得到的其他查询时
```Java exception occurred [class=java.lang.NullPointerException, message=]
Apache.Ignite.Core.Common.IgniteException, Apache.Ignite.Core, Version=2.7.6.40414, Culture=neutral, PublicKeyToken=a487a7ff0b2aaa4a
java.lang.NullPointerException
at org.apache.ignite.internal.processors.cache.IgniteCacheProxyImpl.getConfiguration(IgniteCacheProxyImpl.java:254)
at org.apache.ignite.internal.processors.cache.GatewayProtectedCacheProxy.getConfiguration(GatewayProtectedCacheProxy.java:121)```
对于这两个错误，我尝试通过调用 InitializeCache() 来刷新 IIGnite 来恢复，但我得到了新的错误，例如java.lang.IllegalStateException: Grid is in invalid state to perform this operation. It either not started yet or has already being or have stopped [igniteInstanceName=Dydx.Logistics.Service, state=STOPPED]The issue here is not my appdomain restarting or other similar issues that I have seen in other StackOverflow post. 我不断地运行我的单元测试，直到缓存失败。
请注意，我不在 PROD 环境中的 IIS 中托管 Ignite。但是这个代码对于 DEV 环境来说很容易。希望有人能在这里指出我正确的方向......
异常详情：
EXCEPTION #1A - Occurs when calling DataCache.ContainsKey( new Tuple( somestring, somebyte ) )
Failed to serialize object [typeName=String]
at Apache.Ignite.Core.Impl.PlatformJniTarget.InStreamOutLong[TR](Int32 type, Action`1 outAction, Func`3 inAction, Func`2 readErrorAction)
at Apache.Ignite.Core.Impl.PlatformTargetAdapter.DoOutInOpX(Int32 type, Action`1 outAction, Func`2 inErrorAction)
INNER EXCEPTION
{"class org.apache.ignite.binary.BinaryObjectException: Failed to serialize object [typeName=String]\r\n\tat org.apache.ignite.internal.binary.BinaryClassDescriptor.write(BinaryClassDescriptor.java:840)\r\n\tat org.apache.ignite.internal.binary.BinaryWriterExImpl.marshal0(BinaryWriterExImpl.java:223)\r\n\tat org.apache.ignite.internal.binary.BinaryWriterExImpl.marshal(BinaryWriterExImpl.java:164)\r\n\tat org.apache.ignite.internal.binary.BinaryWriterExImpl.marshal(BinaryWriterExImpl.java:151)\r\n\tat org.apache.ignite.internal.binary.BinaryWriterExImpl.writeObjectDetached(BinaryWriterExImpl.java:1506)\r\n\tat org.apache.ignite.internal.processors.platform.utils.PlatformUtils.writeError(PlatformUtils.java:647)\r\n\tat org.apache.ignite.internal.processors.platform.cache.PlatformCache.processInStreamOutLong(PlatformCache.java:771)\r\n\tat org.apache.ignite.internal.processors.platform.PlatformTargetProxyImpl.inStreamOutLong(PlatformTargetProxyImpl.java:67)\r\nCaused by: class org.apache.ignite.IgniteException: Failed to execute native callback because grid is stopping.\r\n\tat org.apache.ignite.internal.processors.platform.callback.PlatformCallbackGateway.enter(PlatformCallbackGateway.java:1200)\r\n\tat org.apache.ignite.internal.processors.platform.callback.PlatformCallbackGateway.memoryReallocate(PlatformCallbackGateway.java:837)\r\n\tat org.apache.ignite.internal.processors.platform.memory.PlatformExternalMemory.reallocate(PlatformExternalMemory.java:48)\r\n\tat org.apache.ignite.internal.processors.platform.memory.PlatformOutputStreamImpl.ensureCapacity(PlatformOutputStreamImpl.java:305)\r\n\tat org.apache.ignite.internal.processors.platform.memory.PlatformOutputStreamImpl.copyAndShift(PlatformOutputStreamImpl.java:331)\r\n\tat org.apache.ignite.internal.processors.platform.memory.PlatformOutputStreamImpl.writeByteArray(PlatformOutputStreamImpl.java:59)\r\n\tat org.apache.ignite.internal.binary.BinaryWriterExImpl.doWriteString(BinaryWriterExImpl.java:443)\r\n\tat org.apache.ignite.internal.binary.BinaryClassDescriptor.write(BinaryClassDescriptor.java:609)\r\n\t... 7 more\r\n"}
Inspecting the Cache object in the locals window just before where the exception is thrown shows that the Cache is already stopped. When I try to expand the data in the locals window I see:
EXCEPTION #1B - class org.apache.ignite.internal.processors.cache.CacheStoppedException: Failed to perform cache operation (cache is stopped): DataCache
at Apache.Ignite.Core.Impl.PlatformJniTarget.OutObjectInternal(Int32 type)
at Apache.Ignite.Core.Impl.Cache.CacheImpl`2.CreateEnumerator(Boolean loc, Int32 peekModes)
at Apache.Ignite.Core.Impl.Cache.CacheImpl`2.GetEnumerator()
at System.Linq.SystemCore_EnumerableDebugView`1.get_Items()
INNER EXCEPTION:
java.lang.IllegalStateException: class org.apache.ignite.internal.processors.cache.CacheStoppedException: Failed to perform cache operation (cache is stopped): DataCache
at org.apache.ignite.internal.processors.cache.GridCacheGateway.enter(GridCacheGateway.java:164)
at org.apache.ignite.internal.processors.cache.GatewayProtectedCacheProxy.onEnter(GatewayProtectedCacheProxy.java:1555)
at org.apache.ignite.internal.processors.cache.GatewayProtectedCacheProxy.iterator(GatewayProtectedCacheProxy.java:1326)
at org.apache.ignite.internal.processors.platform.cache.PlatformCache.processOutObject(PlatformCache.java:1046)
at org.apache.ignite.internal.processors.platform.PlatformTargetProxyImpl.outObject(PlatformTargetProxyImpl.java:105)
Caused by: class org.apache.ignite.internal.processors.cache.CacheStoppedException: Failed to perform cache operation (cache is stopped): DataCache
... 5 more
EXCEPTION #2 - Occurs when calling dataCache.AsCacheQueryable().Where( x => x.Name == "Something" ).SingleOrDefault();
Java exception occurred [class=java.lang.NullPointerException, message=]
Apache.Ignite.Core.Common.IgniteException, Apache.Ignite.Core, Version=2.7.6.40414, Culture=neutral, PublicKeyToken=a487a7ff0b2aaa4a
java.lang.NullPointerException
at org.apache.ignite.internal.processors.cache.IgniteCacheProxyImpl.getConfiguration(IgniteCacheProxyImpl.java:254)
at org.apache.ignite.internal.processors.cache.GatewayProtectedCacheProxy.getConfiguration(GatewayProtectedCacheProxy.java:121)
at org.apache.ignite.internal.processors.platform.cache.PlatformCache.processOutStream(PlatformCache.java:983)
at org.apache.ignite.internal.processors.platform.PlatformTargetProxyImpl.outStream(PlatformTargetProxyImpl.java:93)
Apache.Ignite.Core.Common.JavaException, Apache.Ignite.Core, Version=2.7.6.40414, Culture=neutral, PublicKeyToken=a487a7ff0b2aaa4a
Stack Trace
at Apache.Ignite.Core.Impl.Unmanaged.Jni.Env.ExceptionCheck()
at Apache.Ignite.Core.Impl.Unmanaged.UnmanagedUtils.TargetOutStream(GlobalRef target, Int32 opType, Int64 memPtr)
at Apache.Ignite.Core.Impl.PlatformJniTarget.OutStream[T](Int32 type, Func`2 readAction)
Again, if I inspect the dataCache just before the error point I see java.lang.IllegalStateException: Grid is in invalid state to perform this operation. It either not started yet or has already being or have stopped [igniteInstanceName=Dydx.Logistics.Service, state=STOPPED]
at org.apache.ignite.internal.GridKernalGatewayImpl.illegalState(GridKernalGatewayImpl.java:201)
at org.apache.ignite.internal.GridKernalGatewayImpl.readLock(GridKernalGatewayImpl.java:95)
at org.apache.ignite.internal.IgniteKernal.guard(IgniteKernal.java:3886)
at org.apache.ignite.internal.IgniteKernal.getOrCreateCache0(IgniteKernal.java:3002)
at org.apache.ignite.internal.IgniteKernal.getOrCreateCache(IgniteKernal.java:2992)
at org.apache.ignite.internal.processors.platform.PlatformProcessorImpl.processInStreamOutObject(PlatformProcessorImpl.java:566)
at org.apache.ignite.internal.processors.platform.PlatformTargetProxyImpl.inStreamOutObject(PlatformTargetProxyImpl.java:79)