Wideclient stops responding

[Red5ive]

Hi Pete,

I have recently started using Widefs to connect PF3-ATC to FSX over my network. Unfortunately I have an intermittent problem when wideclient stops responding. This usually occurs approx 30 mins after the original connection.

Wideserver.log indicates "Error 10053: client socket disconnected at client"

Any ideas would be appreciated.

WideClient.log

WideServer.log

FSUIPC4.ini

[Pete Dowson]

Something is not right asbout the logs, or at least they don't illustrate what you say. Look:

Date (dmy): 26/07/16, Time 09:33:01.035: Server name is FLIGHTSIM
...
  1881653 Restarting service due to zero reception!
  2407876 Incoming connection Accepted ok (skt=5228) TCP
  2407985 Connected to computer "DESKTOP-VU4P01J" running WideClient version 6.999 (skt=5228) TCP
  2861293 Error 10053: client socket disconnected at Client: removing (skt=5228) TCP

The Server starts at 09:33 whilst the Client not till 10:14, about 40 minutes later -- so if this Client log is the one related to the above Server log, the connection shown at 2407985 (i.e. after 40 minutes)  is the FIRST one.

You only ran the Client for less than 8 minutes, closing it after 457.688 seconds (i.e. at about 10:22). The "Error 10053" you see then is simply because WideClient was closed. It is not a network error as such.

However, though you certainly didn't run it for the 30 minutes you stated, this part of the Client log is also relevant:

   457688 Reception maximum:  30 frames/sec, 1782 bytes/sec
   457688 Reception average whilst connected:  29 frames/sec, 1608 bytes/sec
   457688 Transmission maximum:  1 frames/sec, 100 bytes/sec
   457688 Transmission average whilst connected:  0 frames/sec, 20 bytes/sec

So the Client is receiving data from the Server okay, at a decent frame rate too, but very little being sent back to it. Perhaps PF3 doesn't need to send anything? The 1 frame/sec upload is probably just the regular message WideClient sends to the Server to keep it aware it is still there.

So, summarising, I'm not really seeing a problem indicated in the logs at all. Sorry.

Anyway, just in case you are truly experiencing a problem and simply sent the wrong logs, here's some notes. These are in addition to or probably duplicate those in the WideFS manuals, which I hope you referred to first:

First off, you ARE using a wired connection, not wireless, aren't you? Wireless is certainly problematic with the type of traffic WideFs generates normally.

Unfortunately, the Windows network interface doesn't tell applications about retries and errors if it recovers, but if it's got a bad connection somehow (eg. bad Network port on either PC), then many retries could certainly reduce traffic almost to a standstill. Not sure why that would get worse over time, though, unless it is building up a big backlog. If you have alternative Ethernet sockets at either end, try different ones.

You could try setting the protocol in the WideClient INI to UDP, as that has no error correction. If things are going bad then this might give corrupted data which could crash PF3 or even FS, but it would be worth a test to see what the counts were then like. 

Finally, if PF3 really is working ok for those 30 minutes then you could use this desperate measure (from the WideFS Technical manual):

Quote

ReconnectMinutes=0: This parameter can be used to tell WideClient to automatically re-make its connection with the Server at intervals. The value 0 turns this off. Some users find that after continual use for hours the Networking in the client seems to slow down or even give “Bad Data” errors in Clients. There’s no errors actually signalled by Wndows to WideClient and I suspect the problems are down to driver or hardware problems, but the client applications seems to suffer without WideClinet being aware of
this.

If this happens, and it has only been reported when doing flights for four hours or more, try setting a value here – e.g. 240 for reconnection every four hours.

Pete

 

[Red5ive]

On 29/07/2016 at 11:15 AM, Pete Dowson said:

Something is not right asbout the logs, or at least they don't illustrate what you say. Look:

Date (dmy): 26/07/16, Time 09:33:01.035: Server name is FLIGHTSIM
...
  1881653 Restarting service due to zero reception!
  2407876 Incoming connection Accepted ok (skt=5228) TCP
  2407985 Connected to computer "DESKTOP-VU4P01J" running WideClient version 6.999 (skt=5228) TCP
  2861293 Error 10053: client socket disconnected at Client: removing (skt=5228) TCP

The Server starts at 09:33 whilst the Client not till 10:14, about 40 minutes later -- so if this Client log is the one related to the above Server log, the connection shown at 2407985 (i.e. after 40 minutes)  is the FIRST one.

You only ran the Client for less than 8 minutes, closing it after 457.688 seconds (i.e. at about 10:22). The "Error 10053" you see then is simply because WideClient was closed. It is not a network error as such.

However, though you certainly didn't run it for the 30 minutes you stated, this part of the Client log is also relevant:

   457688 Reception maximum:  30 frames/sec, 1782 bytes/sec
   457688 Reception average whilst connected:  29 frames/sec, 1608 bytes/sec
   457688 Transmission maximum:  1 frames/sec, 100 bytes/sec
   457688 Transmission average whilst connected:  0 frames/sec, 20 bytes/sec

So the Client is receiving data from the Server okay, at a decent frame rate too, but very little being sent back to it. Perhaps PF3 doesn't need to send anything? The 1 frame/sec upload is probably just the regular message WideClient sends to the Server to keep it aware it is still there.

So, summarising, I'm not really seeing a problem indicated in the logs at all. Sorry.

Anyway, just in case you are truly experiencing a problem and simply sent the wrong logs, here's some notes. These are in addition to or probably duplicate those in the WideFS manuals, which I hope you referred to first:

First off, you ARE using a wired connection, not wireless, aren't you? Wireless is certainly problematic with the type of traffic WideFs generates normally.

Unfortunately, the Windows network interface doesn't tell applications about retries and errors if it recovers, but if it's got a bad connection somehow (eg. bad Network port on either PC), then many retries could certainly reduce traffic almost to a standstill. Not sure why that would get worse over time, though, unless it is building up a big backlog. If you have alternative Ethernet sockets at either end, try different ones.

You could try setting the protocol in the WideClient INI to UDP, as that has no error correction. If things are going bad then this might give corrupted data which could crash PF3 or even FS, but it would be worth a test to see what the counts were then like. 

Finally, if PF3 really is working ok for those 30 minutes then you could use this desperate measure (from the WideFS Technical manual):

Pete

 

Hi Pete,

Thanks for the prompt reply.  Firstly, I can confirm that I am using a wired connection.  There are two other logs (see attched) which may be relevant.

I think the conflict with the logs may be due to the fact that I aborted my first session because Active Sky Next which I was also running on the client PC, failed to connect to FSX via simconnect.  This I assume explains the Error 10053.

I will try setting the protocol in the WideClient ini to UDP as you suggest.

As it is early days with this setup and the problem is intermitent, I will keep trying, to see if I can obtain more information.

Dave

WideServer0.log

WideServer0.log

[Pete Dowson]

42 minutes ago, Red5ive said:

There are two other logs (see attched) which may be relevant.

They are both the exact same log, in fact, and show nothing useful. Looks like you closed the FS session after about 3 minutes.

Active Sky Next which I was also running on the client PC, failed to connect to FSX via simconnect

Well, SimConnect not connecting might be related.

Pete

 

[Pete Dowson]

43 minutes ago, Red5ive said:

This I assume explains the Error 10053.

The Server never knows why a Client disconnects. The error is simply returned by Windows to tell it the client has gone. In that sense it is only an error if you didn't close it deliberately.

Pete

 

[Red5ive]

Hi Pete,

I have had several more attempts to run PF3 over the network and each time Wideclient stops responding after 5 or 10 mins.  I attach the latest log files, also please see the text below which I have cut and pasted from the Windows Reliabilty Monitor relating to the incident.  I would also point out that no AV and firewall software was enabled during the session.  Any comments would be much appreciated.

Description
Faulting Application Path:    D:\WideClient\WideClient.exe

Problem signature
Problem Event Name:    APPCRASH
Application Name:    WideClient.exe
Application Version:    6.9.9.9
Application Timestamp:    571f8dea
Fault Module Name:    ntdll.dll
Fault Module Version:    10.0.10586.306
Fault Module Timestamp:    571afb7f
Exception Code:    c0000005
Exception Offset:    0003c56e
OS Version:    10.0.10586.2.0.0.768.101
Locale ID:    2057
Additional Information 1:    bc98
Additional Information 2:    bc988609915e0d2320e58b285b07f7a9
Additional Information 3:    b603
Additional Information 4:    b6037cec1a1e7a676b742f3c16e92722

Extra information about the problem
Bucket ID:    af21370777cf1688d4d7a0f7a791a785 (108227224329)

Regards

WideClient.log

WideServer.log

[Pete Dowson]

Sorry, there's no useful information here. Maybe a debug log with WideClient would show more, but beware the size of the log.  You will need to ZIP it and send it to me as an email attachment -- petedowson@btconnect.com. Debug options are shown in the Technical guide for WideFS -- try Log=All first, then Log=DebugAll (separate sessions). Send e both in suitably named ZIP files.

Incidentally, have you checked that PF3 will actually run Networked? Are others doing it? Is it documented?

Pete

 

[Red5ive]

I Thought I ought to post a note in oder to close this issue.

As you found no issues with the log files, this prompted me to dig deeper into my system. This culminated in me carrying out a system file check by running "sfc /scannow" via the Command Prompt. The scan found corrupt files which it was unable to fix.  I hadn't expected this as the computer was fairly new and I had only installed a small number of software packages, ASNext, PF3 and WideFS.

To cut a long storey short, I reset the computer to the factory settings, re-installed the software and am now able to run PF3 over the network using WideFS without any issues.

Sorry for having wasted your time with this issue.

Regards Dave