Agents have finished their builds, but the server is still waiting; can't kill the service either. How do I troubleshoot this?

This may be related to TeamCity 8.1.2 (Windows server 2008 R2); once in a while, TeamCity server cannot seem to get build info from the agents.  There are no commonalities

Symptoms:

1.  TeamCity web will show that a build is going on and on on the agent; however on the agents, I look at the logs and the build has finished! But the server is still waiting and shows that the build is now going into overtime.  The logs are partially completed and you can see more output is expected but nothing appears.

2.  Cannot kill the TeamCity service.  Normally I would restart the service but the service hangs as well, in the 'stopping' state.  I have to go into Process Explorer and kill the java processes myself.

On the agent logs (teamcity-agent.log) I see this repeating itself

[2014-05-07 11:12:52,091]   INFO - r.agent.impl.AgentLogProxyImpl - Ping OK and registered on the server again: 32894
[2014-05-07 11:12:52,091]   INFO - r.agent.impl.AgentLogProxyImpl - Pending remote commands count: 11
[2014-05-07 11:12:57,521]   INFO -    jetbrains.buildServer.AGENT - Done: Pre-processing artifacts if needed (archiving)
[2014-05-07 11:12:57,521]   INFO - ages.BuildFinishStagesExecutor - Call finish stage jetbrains.buildServer.agent.impl.buildStages.finishStages.WaitForArtifactsPublishedFStage
[2014-05-07 11:17:52,305]   INFO - r.agent.impl.AgentLogProxyImpl - Failed to perform remote command 100 log messages for build 32894, error: jetbrains.buildServer.xmlrpc.RemoteCallException: Call http://TCServer.internaldomain/RPC2 buildServer.log: org.apache.xmlrpc.XmlRpcException: jetbrains.buildServer.TeamCityRuntimeException: Timeout exceeded while waiting for messages processing
[2014-05-07 11:17:52,305]   INFO - r.agent.impl.AgentLogProxyImpl - Adding ping and re-register command on top of the remote command queue
[2014-05-07 11:17:57,539]   INFO - buildServer.AGENT.registration - Registering on server http://TCServer.internaldomain, AgentDetails{Name='TCAgent.internaldomain', AgentId=6, BuildId=32894, AgentOwnAddress='null', AlternativeAddresses=[10.55.109.166], Port=81, Version='29993', PluginsVersion='29993-md5-6046f906792a5ced1814676fc926e712', AvailableRunners=[Ant, dotnet-tools-dupfinder, dotnet-tools-inspectcode, Duplicator, FxCop, gradle-runner, Inspection, Ipr, jb.nuget.installer, jb.nuget.pack, jb.nuget.publish, jetbrains.dotNetGenericRunner, jetbrains.mspec, jetbrains_powershell, JPS, Maven2, MSBuild, MSTest, NAnt, NUnit, rake-runner, simpleRunner, sln2003, VS.Solution], AvailableVcs=[perforce, mercurial, jetbrains.git, tfs, svn, cvs], AuthorizationToken='a2a0473441c33f89a9a077afeec31b45', PingCode='c8QA4wCUP7tLfMwIO43eagdlkU2ne1wv'}
[2014-05-07 11:17:57,760]   INFO - buildServer.AGENT.registration - Server confirmed we are still registered
[2014-05-07 11:17:57,761]   INFO - r.agent.impl.AgentLogProxyImpl - Ping OK and registered on the server again: 32894
[2014-05-07 11:17:57,761]   INFO - r.agent.impl.AgentLogProxyImpl - Pending remote commands count: 21
[2014-05-07 11:22:57,947]   INFO - r.agent.impl.AgentLogProxyImpl - Failed to perform remote command 100 log messages for build 32894, error: jetbrains.buildServer.xmlrpc.RemoteCallException: Call http://TCServer.internaldomain/RPC2 buildServer.log: org.apache.xmlrpc.XmlRpcException: jetbrains.buildServer.TeamCityRuntimeException: Timeout exceeded while waiting for messages processing
[2014-05-07 11:22:57,948]   INFO - r.agent.impl.AgentLogProxyImpl - Adding ping and re-register command on top of the remote command queue
[2014-05-07 11:23:03,161]   INFO - buildServer.AGENT.registration - Registering on server http://TCServer.internaldomain, AgentDetails{Name='TCAgent.internaldomain', AgentId=6, BuildId=32894, AgentOwnAddress='null', AlternativeAddresses=[10.55.109.166], Port=81, Version='29993', PluginsVersion='29993-md5-6046f906792a5ced1814676fc926e712', AvailableRunners=[Ant, dotnet-tools-dupfinder, dotnet-tools-inspectcode, Duplicator, FxCop, gradle-runner, Inspection, Ipr, jb.nuget.installer, jb.nuget.pack, jb.nuget.publish, jetbrains.dotNetGenericRunner, jetbrains.mspec, jetbrains_powershell, JPS, Maven2, MSBuild, MSTest, NAnt, NUnit, rake-runner, simpleRunner, sln2003, VS.Solution], AvailableVcs=[perforce, mercurial, jetbrains.git, tfs, svn, cvs], AuthorizationToken='a2a0473441c33f89a9a077afeec31b45', PingCode='c8QA4wCUP7tLfMwIO43eagdlkU2ne1wv'}
[2014-05-07 11:23:03,380]   INFO - buildServer.AGENT.registration - Server confirmed we are still registered
[2014-05-07 11:23:03,381]   INFO - r.agent.impl.AgentLogProxyImpl - Ping OK and registered on the server again: 32894
[2014-05-07 11:23:03,381]   INFO - r.agent.impl.AgentLogProxyImpl - Pending remote commands count: 21
[2014-05-07 11:28:03,563]   INFO - r.agent.impl.AgentLogProxyImpl - Failed to perform remote command 100 log messages for build 32894, error: jetbrains.buildServer.xmlrpc.RemoteCallException: Call http://TCServer.internaldomain/RPC2 buildServer.log: org.apache.xmlrpc.XmlRpcException: jetbrains.buildServer.TeamCityRuntimeException: Timeout exceeded while waiting for messages processing
[2014-05-07 11:28:03,563]   INFO - r.agent.impl.AgentLogProxyImpl - Adding ping and re-register command on top of the remote command queue
[2014-05-07 11:28:08,776]   INFO - buildServer.AGENT.registration - Registering on server http://TCServer.internaldomain, AgentDetails{Name='TCAgent.internaldomain', AgentId=6, BuildId=32894, AgentOwnAddress='null', AlternativeAddresses=[10.55.109.166], Port=81, Version='29993', PluginsVersion='29993-md5-6046f906792a5ced1814676fc926e712', AvailableRunners=[Ant, dotnet-tools-dupfinder, dotnet-tools-inspectcode, Duplicator, FxCop, gradle-runner, Inspection, Ipr, jb.nuget.installer, jb.nuget.pack, jb.nuget.publish, jetbrains.dotNetGenericRunner, jetbrains.mspec, jetbrains_powershell, JPS, Maven2, MSBuild, MSTest, NAnt, NUnit, rake-runner, simpleRunner, sln2003, VS.Solution], AvailableVcs=[perforce, mercurial, jetbrains.git, tfs, svn, cvs], AuthorizationToken='a2a0473441c33f89a9a077afeec31b45', PingCode='c8QA4wCUP7tLfMwIO43eagdlkU2ne1wv'}
[2014-05-07 11:28:08,996]   INFO - buildServer.AGENT.registration - Server confirmed we are still registered
[2014-05-07 11:28:08,996]   INFO - r.agent.impl.AgentLogProxyImpl - Ping OK and registered on the server again: 32894
[2014-05-07 11:28:08,996]   INFO - r.agent.impl.AgentLogProxyImpl - Pending remote commands count: 21

This is on all agent logs, even the agent which is on the same machine as the server.  

On the TeamCity server itself, I looked at "teamcity-server.log" but I don't see anything unusual at those times.  Just a few of these

[2014-05-07 11:16:48,183]   INFO -   jetbrains.buildServer.SERVER - Agent description updated for agent: TCAgent {id=1, host=127.0.0.1:9090, agentTypeId=1, pool=Default}
[2014-05-07 11:17:06,750]   INFO -   jetbrains.buildServer.SERVER - Agent description updated for agent: TCAgent {id=1, host=127.0.0.1:9090, agentTypeId=1, pool=Default}
[2014-05-07 11:18:01,743]   INFO -   jetbrains.buildServer.SERVER - Agent description updated for agent: TCAgent {id=1, host=127.0.0.1:9090, agentTypeId=1, pool=Default}
[2014-05-07 11:19:04,637]   INFO -   jetbrains.buildServer.SERVER - Agent description updated for agent: TCAgent2 {id=5, host=10.55.109.167:8090, agentTypeId=5, pool=Default}
[2014-05-07 11:19:21,088]   INFO -   jetbrains.buildServer.SERVER - Agent description updated for agent: TCAgent2 {id=5, host=10.55.109.167:8090, agentTypeId=5, pool=Default}
[2014-05-07 11:19:31,857]   INFO -   jetbrains.buildServer.SERVER - Agent description updated for agent: TCAgent2 {id=5, host=10.55.109.167:8090, agentTypeId=5, pool=Default}
[2014-05-07 11:20:03,275]   INFO -   jetbrains.buildServer.SERVER - Agent description updated for agent: TCAgent2 {id=5, host=10.55.109.167:8090, agentTypeId=5, pool=Default}

I am a bit stuck because I cannot see what is wrong, but these hangups seem to be happening randomly.  We've only noticed this problem since upgrading to 8.1.2 but don't know if it's related specifically to this version.  

Is there anywhere else I could be looking? Any suggestions on how to fix this would also be helpful

Thanks