| 2013-07-25 17:11:59 |
Oleg Gerovich |
description |
I have the following test:
using Microsoft.Hpc.Scheduler;
using NUnit.Framework;
namespace TestHPC
{
[TestFixture]
internal class TestHpc
{
[TestCase("HEAD-NODE-SERVER", 9)]
public static void AllCoresPresent(string headNode, int expectedCores)
{
int totalCores = 0;
using (IScheduler scheduler = new Scheduler())
{
scheduler.Connect(headNode);
ISchedulerCollection nodes = scheduler.GetNodeList(null, null);
foreach (ISchedulerNode node in nodes)
{
totalCores += node.NumberOfCores;
}
}
Assert.AreEqual(expectedCores, totalCores);
}
}
}
I am using a Visual Studio 2010 solution that has the project with this test build against .NET 4.0. This is the only test in the solution.
If I use NUnit 2.6.0 GUI or console runner, the test hangs in the middle.
If I use ReSharper 7.1.1 Unit Test runner pointed at my local NUnit 2.6.0 and run the test from the project compiled against 2.6.0 nunit.framework.dll, the test passes quite quickly.
If I switch ReSharper to run tests with build-in NUnit 2.6.2 and compile against 2.6.2, the test fails with this error:
Microsoft.Hpc.Scheduler.Properties.SchedulerException : No connection could be made because the target machine actively refused it 127.0.0.1:6729
----> System.Net.Sockets.SocketException : No connection could be made because the target machine actively refused it
Server stack trace:
at Microsoft.Hpc.Scheduler.NodeManagement.NodeQuery.Reconnect()
at Microsoft.Hpc.Scheduler.NodeManagement.NodeQuery.HandleException(Exception e)
at Microsoft.Hpc.Scheduler.NodeManagement.NodeQuery.EnumerateGroupsForNode(String nodeName)
at Microsoft.Hpc.Scheduler.Store.NodePropertyHandler.Tags.GetPropFromQuery(QueryContextBase ctx, PropertyId pid, ref StoreProperty prop)
at Microsoft.Hpc.Scheduler.Store.QueryContextBase.GetPropFromQuery(PropertyId pid)
at Microsoft.Hpc.Scheduler.Store.QueryContextBase.ExecuteRowSetQuery(List`1 ids, PropertyId[] pids)
at Microsoft.Hpc.Scheduler.Store.SnapShotRowSet.GetData(Int32 firstRow, Int32 lastRow, ref Int32 rowCount)
at Microsoft.Hpc.Scheduler.Store.SchedulerStoreInternal.RowSet_GetData(ref ConnectionToken token, Int32 rowsetId, Int32 firstRow, Int32 lastRow)
at Microsoft.Hpc.Scheduler.Store.SchedulerStoreInternal.RowEnum_GetRows(ref ConnectionToken token, Int32 id, Int32 numberOfRows)
at System.Runtime.Remoting.Messaging.StackBuilderSink._PrivateProcessMessage(IntPtr md, Object[] args, Object server, Int32 methodPtr, Boolean fExecuteInContext, Object[]& outArgs)
at System.Runtime.Remoting.Messaging.StackBuilderSink.SyncProcessMessage(IMessage msg, Int32 methodPtr, Boolean fExecuteInContext)
Exception rethrown at [0]:
at Microsoft.Hpc.Scheduler.Store.StoreServer.HandleException(Exception e, Boolean asyncReconnect)
at Microsoft.Hpc.Scheduler.Store.StoreServer.RowEnum_GetRows(Int32 id, Int32 numberOfRows)
at Microsoft.Hpc.Scheduler.Store.LocalRowEnumerator.GetRows2(Int32 numberOfRows)
at Microsoft.Hpc.Scheduler.Store.RowEnumeratorEnumerator.MoveNext()
at Microsoft.Hpc.Scheduler.Scheduler.GetNodeList(IFilterCollection filter, ISortCollection sort)
at TestHPC.TestHpc.AllCoresPresent(String headNode, Int32 expectedCores) in TestHPC.cs: line 18
--SocketException
Server stack trace:
at System.Net.Sockets.Socket.DoConnect(EndPoint endPointSnapshot, SocketAddress socketAddress)
at System.Net.Sockets.Socket.Connect(EndPoint remoteEP)
at System.Runtime.Remoting.Channels.RemoteConnection.CreateNewSocket(EndPoint ipEndPoint)
at System.Runtime.Remoting.Channels.RemoteConnection.CreateNewSocket()
at System.Runtime.Remoting.Channels.SocketCache.GetSocket(String machinePortAndSid, Boolean openNew)
at System.Runtime.Remoting.Channels.Tcp.TcpClientTransportSink.SendRequestWithRetry(IMessage msg, ITransportHeaders requestHeaders, Stream requestStream)
at System.Runtime.Remoting.Channels.Tcp.TcpClientTransportSink.ProcessMessage(IMessage msg, ITransportHeaders requestHeaders, Stream requestStream, ref ITransportHeaders responseHeaders, ref Stream responseStream)
at Microsoft.Hpc.ClientSink.ProcessMessage(IMessage message, ITransportHeaders requestHeaders, Stream requestStream, ref ITransportHeaders responseHeaders, ref Stream responseStream)
at System.Runtime.Remoting.Channels.BinaryClientFormatterSink.SyncProcessMessage(IMessage msg)
Exception rethrown at [0]:
at System.Runtime.Remoting.Proxies.RealProxy.HandleReturnMessage(IMessage reqMsg, IMessage retMsg)
at System.Runtime.Remoting.Proxies.RealProxy.PrivateInvoke(ref MessageData msgData, Int32 type)
at Microsoft.Hpc.Scheduler.NodeManagement.INodeQuery.EnumerateNodeGroups()
at Microsoft.Hpc.Scheduler.NodeManagement.NodeQuery.Reconnect()
If I use NUnit 2.6.2 GUI or console runner, the test hangs in the middle.
There are no firewalls blocking port 6729 on the HPC server. All I did was switch the version of NUnit.
There seems to be a difference in how ReSharper is launching NUnit from the way NUnit GUI and console runner work. If I did not have ReSharper, I would have no clue that my test could pass.
ReSharper's behavior is observed in TeamCity when I run the test using NUnit runner. Looks like JetBrains have some consistency across their products that work better for execution of NUnit tests than native tools.
I have a variety of other tests that work just fine after the switch, so the problem is constrained to Microsoft.Hpc.Scheduler tests (for now). I can provide more information if needed. Please help! |
I have the following test:
using Microsoft.Hpc.Scheduler;
using NUnit.Framework;
namespace TestHPC
{
[TestFixture]
internal class TestHpc
{
[TestCase("HEAD-NODE-SERVER", 9)]
public static void AllCoresPresent(string headNode, int expectedCores)
{
int totalCores = 0;
using (IScheduler scheduler = new Scheduler())
{
scheduler.Connect(headNode);
ISchedulerCollection nodes = scheduler.GetNodeList(null, null);
foreach (ISchedulerNode node in nodes)
{
totalCores += node.NumberOfCores;
}
}
Assert.AreEqual(expectedCores, totalCores);
}
}
}
I am using a Visual Studio 2010 solution that has the project with this test build against .NET 4.0. This is the only test in the solution.
If I use NUnit 2.6.0 GUI or console runner, the test hangs in the middle.
If I use ReSharper 7.1.1 Unit Test runner pointed at my local NUnit 2.6.0 and run the test from the project compiled against 2.6.0 nunit.framework.dll, the test passes quite quickly.
If I switch ReSharper to run tests with built-in NUnit 2.6.2 and compile against 2.6.2, the test fails with this error:
Microsoft.Hpc.Scheduler.Properties.SchedulerException : No connection could be made because the target machine actively refused it 127.0.0.1:6729
----> System.Net.Sockets.SocketException : No connection could be made because the target machine actively refused it
Server stack trace:
at Microsoft.Hpc.Scheduler.NodeManagement.NodeQuery.Reconnect()
at Microsoft.Hpc.Scheduler.NodeManagement.NodeQuery.HandleException(Exception e)
at Microsoft.Hpc.Scheduler.NodeManagement.NodeQuery.EnumerateGroupsForNode(String nodeName)
at Microsoft.Hpc.Scheduler.Store.NodePropertyHandler.Tags.GetPropFromQuery(QueryContextBase ctx, PropertyId pid, ref StoreProperty prop)
at Microsoft.Hpc.Scheduler.Store.QueryContextBase.GetPropFromQuery(PropertyId pid)
at Microsoft.Hpc.Scheduler.Store.QueryContextBase.ExecuteRowSetQuery(List`1 ids, PropertyId[] pids)
at Microsoft.Hpc.Scheduler.Store.SnapShotRowSet.GetData(Int32 firstRow, Int32 lastRow, ref Int32 rowCount)
at Microsoft.Hpc.Scheduler.Store.SchedulerStoreInternal.RowSet_GetData(ref ConnectionToken token, Int32 rowsetId, Int32 firstRow, Int32 lastRow)
at Microsoft.Hpc.Scheduler.Store.SchedulerStoreInternal.RowEnum_GetRows(ref ConnectionToken token, Int32 id, Int32 numberOfRows)
at System.Runtime.Remoting.Messaging.StackBuilderSink._PrivateProcessMessage(IntPtr md, Object[] args, Object server, Int32 methodPtr, Boolean fExecuteInContext, Object[]& outArgs)
at System.Runtime.Remoting.Messaging.StackBuilderSink.SyncProcessMessage(IMessage msg, Int32 methodPtr, Boolean fExecuteInContext)
Exception rethrown at [0]:
at Microsoft.Hpc.Scheduler.Store.StoreServer.HandleException(Exception e, Boolean asyncReconnect)
at Microsoft.Hpc.Scheduler.Store.StoreServer.RowEnum_GetRows(Int32 id, Int32 numberOfRows)
at Microsoft.Hpc.Scheduler.Store.LocalRowEnumerator.GetRows2(Int32 numberOfRows)
at Microsoft.Hpc.Scheduler.Store.RowEnumeratorEnumerator.MoveNext()
at Microsoft.Hpc.Scheduler.Scheduler.GetNodeList(IFilterCollection filter, ISortCollection sort)
at TestHPC.TestHpc.AllCoresPresent(String headNode, Int32 expectedCores) in TestHPC.cs: line 18
--SocketException
Server stack trace:
at System.Net.Sockets.Socket.DoConnect(EndPoint endPointSnapshot, SocketAddress socketAddress)
at System.Net.Sockets.Socket.Connect(EndPoint remoteEP)
at System.Runtime.Remoting.Channels.RemoteConnection.CreateNewSocket(EndPoint ipEndPoint)
at System.Runtime.Remoting.Channels.RemoteConnection.CreateNewSocket()
at System.Runtime.Remoting.Channels.SocketCache.GetSocket(String machinePortAndSid, Boolean openNew)
at System.Runtime.Remoting.Channels.Tcp.TcpClientTransportSink.SendRequestWithRetry(IMessage msg, ITransportHeaders requestHeaders, Stream requestStream)
at System.Runtime.Remoting.Channels.Tcp.TcpClientTransportSink.ProcessMessage(IMessage msg, ITransportHeaders requestHeaders, Stream requestStream, ref ITransportHeaders responseHeaders, ref Stream responseStream)
at Microsoft.Hpc.ClientSink.ProcessMessage(IMessage message, ITransportHeaders requestHeaders, Stream requestStream, ref ITransportHeaders responseHeaders, ref Stream responseStream)
at System.Runtime.Remoting.Channels.BinaryClientFormatterSink.SyncProcessMessage(IMessage msg)
Exception rethrown at [0]:
at System.Runtime.Remoting.Proxies.RealProxy.HandleReturnMessage(IMessage reqMsg, IMessage retMsg)
at System.Runtime.Remoting.Proxies.RealProxy.PrivateInvoke(ref MessageData msgData, Int32 type)
at Microsoft.Hpc.Scheduler.NodeManagement.INodeQuery.EnumerateNodeGroups()
at Microsoft.Hpc.Scheduler.NodeManagement.NodeQuery.Reconnect()
If I use NUnit 2.6.2 GUI or console runner, the test hangs in the middle.
There are no firewalls blocking port 6729 on the HPC server. All I did was switch the version of NUnit.
There seems to be a difference in how ReSharper is launching NUnit from the way NUnit GUI and console runner work. If I did not have ReSharper, I would have no clue that my test could pass.
ReSharper's behavior is observed in TeamCity when I run the test using NUnit runner. Looks like JetBrains have some consistency across their products that work better for execution of NUnit tests than native tools.
I have a variety of other tests that work just fine after the switch, so the problem is constrained to Microsoft.Hpc.Scheduler tests (for now). I can provide more information if needed. Please help! |
|
