PNRP service under Vista Ultimate(RTM) version is very instable !!
I'm on a project about WCF P2P. Under Windows XP(PNRP V2) everything fine. But under Vista Ultimate (RTM), I have to always restart PNRP service manually after some P2P testing.
Anybody knows why?
Hi Hailong,
Sorry to hear you are running into problems. Could you please elaborate on why you need to restart the service ? What kind of environment (internet vs lan; single machine vs multi-machine) are you testing on?
As far as I've tested and worked with PNRP on Vista, I've never run into issues where I've crashed and have had to restart the PNRP service and am curious to know more about the issue you are seeing.
Thanks,
Shalini.
--Please do not send email directly to this alias. This alias is for newsgroup purposes only. This posting is provided "AS IS" with no warranties, and confers no rights.
Hi Shalini,
Thank you ! :)
My testing environment are:
1: Single machine local lan
OS: Windows Vista(TM) Ultimate, Japanese Version with all new updates, no virous protection software
Hardware: DLL dimension 9200( Intel(R) Core(TM)2 6400 @ 2.13GHZ 2.13GHZ; Memory 2046MB )
2: multi-machine local lan
OS: same as above
Hardware 1: same as above
Hardware 2: Dll dimension E521(AMD Athlon(tm) 64x2 Dual Core Processor 3800+ 2.00 GHZ; Memory 958MB)
Error like this:(I'm working for a Japanese company, but I cann't read Japanese :( )
After running my program about 10 times(crash or not), ServiceHost.Open() through a CommunicationException
System.ServiceModel.CommunicationException: ピア リゾルバが例外をスローしました。InnerException を参照してください。 > System.ServiceModel.Channels.PnrpPeerResolver+PnrpException: 到達できないネットワークでソケット操作を実行しようとしました。
If I didn't restart PNRP service after error occured, wait about 20 minutes, everything becomes to be fine.........
There is another problem, Teredo doesn't work as previrous Vista version. I didn't check the detial.
Hi Hailong,
Thanks for the information. One of my colleagues, Peter, was able to provide a rough translation of hte InnerException that Pnrp reported - note that this looks more and more like an underlying Network issue rather than Pnrp. What exactly is causing your network problem is something we need to figure out better:
Here are some troubleshooting steps to begin with:-
1. Change the language settings on the machine you are testing from Japanese to English - I believe this can be done by going to Control Panel => Regional And Language Options. This will help us more in understanding the exact error that you are getting.
2. Try and repro the error and at the time it happens, can you share hte output of "ipconfig /all"
3. From an elevated command window (cmd => Run As => Administrator), the output of "netsh p2p pnrp cl sho state" - this will give a list of all PNRP available clouds (that correspond to Global/LinkLocal/SitelOcal IPv6 addresses for your machine) and the state they are in -
Please share out the output (you can also email me on shalinij@microsoft.com)
Below is the translation:
*******************************************
System.ServiceModel.CommunicationException: ? ?InnerException ? > System.ServiceModel.Channels.PnrpPeerResolver+PnrpException: ???
<Translation of the error message>
Piarizoruba ga <exception> wo threw. Inner exception wo sanshou <consult> shite kudasai. <English> toutatsu <arrive> dekinai network de socket souse <operation, handling, control> wo jikkou <execute> shiyou to shimashita.
Piarizoruba is a mangled English word, (or possibly two together), it is hard to figure out what it might be. Pia might be “peer” or “pair”, not sure about zoruba. Probably “peer”, given the context.
Pia rizoruba threw an exception.
Please see the inner exception. (pnrp exception).
Unable to arrive (unreachable?) network by means of socket operation obj. tried and failed to execute.
Windows tried, but couldn’t reach the network over the socket?
</translation>
Thanks,
Shalini.
I deleted my last reply, because it included some security information, I'm so sorry for that.
I've sent a email to Shalini. If the problem resolved latter, I'll publish the final result here.
Thank Shalini's help very much, below is the explanation and resolution, by Shalini :
Type : client
Server Name : teredo.ipv6.microsoft.com.
Client Refresh Interval : 60 seconds
Client Port : 34567
State : probe(restricted)
Type : teredo client
Network : unmanaged
NAT : restricted
Type : client
Server Name : teredo.ipv6.microsoft.com.
Client Refresh Interval : 60 seconds
Client Port : 34567
State : qualified
Type : teredo client
Network : unmanaged
NAT : restricted
Hi,
I've also found it quite unstable on Vista Ultimate. In my corporate environment, Teredo is not accessable and so is in the offline state. However, the confusing part about using PNRP is that it works fine for testing apps locally and then all of a sudden I get communicationexceptions and have to restart the PNRP service before it starts working again. It works fine running the simple chat sample in the SDK, can run 50 instances on the local machine and they all work fine using the default PNRP resolver, and then randomly later on it stops working. I'm totally new to the peer2peer stuff in the framework.
I've also added the Windows Collaboration Computer Name Registration Service to the firewall without any luck.
Can anybody please explain this behaviour?
The exception is:
System.ServiceModel.CommunicationException: The Peer resolver threw an exception. Please refer to InnerException. > System.ServiceModel.Channels.PnrpPeerResolver+PnrpException: A socket operation was attempted to an unreachable network
at System.ServiceModel.Channels.PeerExceptionHelper.ThrowPnrpError(Int32 errorCode, String cloud, Boolean trace)at System.ServiceModel.Channels.PnrpPeerResolver.UnsafePnrpNativeMethods.PeerNameRegistrar.InvokeService(WsaQuerySet registerQuery, WsaSetServiceOp op, Int32 flags)at System.ServiceModel.Channels.PnrpPeerResolver.UnsafePnrpNativeMethods.PeerNameRegistrar.Register(PnrpRegistration registration, TimeSpan timeout)at System.ServiceModel.Channels.PnrpPeerResolver.Register(String meshId, PeerNodeAddress nodeAddress, TimeSpan timeout)at System.ServiceModel.Channels.PeerNodeImplementation.RegisterAddress(String lclMeshId, PeerNodeAddress nodeAddress, TimeSpan timeout) End of inner exception stack trace
Server stack trace:
at System.ServiceModel.AsyncResult.End[TAsyncResult](IAsyncResult result)
at System.ServiceModel.Channels.PeerNodeImplementation.SimpleStateManager.EndOpen(IAsyncResult result)
at System.ServiceModel.Channels.PeerNodeImplementation.SimpleStateManager.Open(TimeSpan timeout, Boolean waitForOnline)
at System.ServiceModel.Channels.PeerNodeImplementation.Open(TimeSpan timeout, Boolean waitForOnline)
at System.ServiceModel.Channels.PeerDuplexChannel.OnOpen(TimeSpan timeout)
at System.ServiceModel.Channels.CommunicationObject.Open(TimeSpan timeout)
at System.ServiceModel.Channels.ServiceChannel.OnOpen(TimeSpan timeout)
at System.ServiceModel.Channels.CommunicationObject.Open(TimeSpan timeout)
at System.ServiceModel.Channels.CommunicationObject.Open()
Exception rethrown at [0]:
at System.Runtime.Remoting.Proxies.RealProxy.HandleReturnMessage(IMessage reqMsg, IMessage retMsg)
at System.Runtime.Remoting.Proxies.RealProxy.PrivateInvoke(MessageData& msgData, Int32 type)
at System.ServiceModel.ICommunicationObject.Open()
at Microsoft.ServiceModel.Samples.ChatApp.Main() in ...
Thanks
Chris
Hi Chris:
I'm not sure whether we met the same problem. Do your test PC behind a NAT? Here is my suggest:
1: Disable teredo
Run "command prompt" as administrator, type following command "netsh interface teredo set state disable", enter
Test your program again. If the same problem occures, I think the problem is not because of teredo service.
2: Enable teredo
If no problems without teredo service running, enable teredo as following
Run "command prompt" as administrator, type following command "netsh interface teredo set state default", enter
Test your program again. If the same problem occures, please update your OS with "Windows Update". Then test again
Actually, I'm a beginner of WCF p2p too, I know few about it.
Thanks for your help! Yes my PC is behind a NAT. I've tried what you suggested and that seems like its the cause. If its working and then I disable teredo, it can't resolve anymore, if I enable it again, it resolves again. However after a little while teredo goes offline again.
Heres the state of teredo when its working:
netsh interface teredo>show state
Teredo Parameters
Type : default
Server Name : teredo.ipv6.microsoft.com.
Client Refresh Interval : 30 seconds
Client Port : unspecified
State : probe (primary server)
Client Type : teredo client
Network : unmanaged
NAT : none (global connectivity)
and when its not:
netsh interface teredo>show state
Teredo Parameters
Type : default
Server Name : teredo.ipv6.microsoft.com.
Client Refresh Interval : 30 seconds
Client Port : unspecified
State : offline
Error : primary teredo server unreachable over UDP
Therefore the bit that confuses me is that why when you enable it or start the service, it starts up in a kind of connected state, to only find out later that my computer can't work over teredo because of the environment? I would of expected it would start up in an offline state, and only once its determined it can work over my environment it would go into an online state. Therefore I would expect that the chat sample would always not work using PNRP, rather than be up and then down, giving a feeling of instability.
There must be something simple I'm not understanding or overlooking?
Thanks!
Chris
Hi Chris,
What you are seeing is a Teredo "feature" where by it goes offline/dormant after a while - so when the chat sample starts, Teredo is in the "qualified" (online) state and things work - but it is possible that it goes offline, and then the address that Peer channel has been using is no longer valid - and we end up throwing that exception.
Like I suggested to Hailong, it's a good idea to catch CommunicationExceptions in your application to make it robust and resilient to such underlying network conditions/changes. Note that you would need to recreate your client channel (proxy) if you intend to continue to use your application, since we abort the channel upon such exceptions.
I am also looking to see if there are some diagnostic tools that would periodically look at the Teredo supplied Global IPv6 address and make sure it is usable (by pinging the Global microsoft Teredo seed server - (by the way, make sure that the Teredo UDP port is enabled in your firewall). One way is to leave the follwoing netsh context running on a separate command window while running/testing your application:
Admin cmd shell:
"netsh p2p pnrp cl sho list"
Let me know if that helps!
-Shalini.
Hi Shalini,
Thanks for your help, it gives a bit more insight into whats happening.
Yes the application will handle the exceptions and retry connecting. What I was thinking is that the app shows a status icon somewhere on the main screen showing the connectivity. However, I'd like it to not show its connected to the p2p mesh initially if its doomed to go offline a little bit later when it can't send UDP messages to the Teredo seed server. (in the teredo blocked scenario). Is this possible?
Chris
