TELNET CLIENT PROBLEM

Hello Dave Wilson,

You have any news on the problem I mentioned above.

Best regards.

Ciro

I have this problem in my queue to look at but I've not yet had a chance to look into it - sorry. I plan to have this investigated and a fix in place in time for our next StellarisWare release in early August.

Hi Dave,

Ok, I'll be anxiously awaiting the new version.

Tks,

Ciro

Ciro,

  I've started looking into this problem but am having trouble reproducing it on either release 5570 or the latest version. Can you offer any more information on what you are seeing when the problem occurs, how your S2E port is configured and what your networking setup looks like?

  In case it helps, here's what I did to reproduce the problem.

1. Attach S2E to our office LAN and to COM1on my PC.
2. Run TeraTerm and open COM1 using parameters 115200/8/N/1.
3. Started the Windows telnet server on my PC.
4. Run the StellarisWare "finder" app to determine the IP address of the S2E module.
5. Use a web browser to configure Port 0 of the S2E to use Telnet Client mode connecting to my PC's IP address, server port 23, local port 1234, serial parameters as above and 0 in the timeout field.
6. Ran WireShark and captured all telnet traffic between my PC and the S2E module. This shows traffic and I see the welcome banner and "login:" prompt in the TeraTerm window as expected.
7. Stopped the Windows telnet server using "tlntadmn stop". This caused the connection to be shut down. Wireshark shows a constant flood of reconnection attempts from the S2E.
8. Restarted the Windows telnet server using "tlntadmn start". This caused the next S2E reconnection request to succeed and I get the login banner and prompt once again.
9. Removed the Ethernet cable from the S2E. Typed a few characters on TeraTerm then reconnected the cable. The connection was reinitiated successfully (though I didn't check whether any data was lost during the interruption).

  If it would be possible for you to capture a Wireshark trace showing the problem occurring, that would also be enormously useful.

This is helpful - thanks. I realised, of course, that I had not tried Port 1 so I'll look into this now in case it's a port-specific problem. The telnet server I am using is the default Windows XP telnet service. I've not changed any parameters away from the defaults as far as I can remember.

I'll post more once I've had a chance to try Port 1 (after finding some RS232 level shifters since that's port doesn't have 12V transceivers as far as I can remember) and also after trying both ports simultaneously.

Edited Later:

I'm happy to report that I can now reproduce a problem when running 2 telnet sessions simultaneously. I've not captured a trace of exactly what is going on but this is a step forward. Could you please let me know what your S2E settings are for each port so that I can try to duplicate them exactly? So far, I only see this problem if I'm actually logged in to the Windows system then remove the Ethernet connection for some period of time before reconnecting. In this case, one of the two sessions hangs and doesn't reconnect properly even though Windows considers the telnet session active. I can cause a reconnection by killing the session from the server end (tlntadmn -k <session #>).

From what we figure out, killing the server won't be usefull, we were able to trace the problem and one of the S2E clients just "die" and stops requesting connection after a while (if the server is down and doesn't respond) . follow a snap shot from the wireshark log with remarks.

       If you find usefull please give us an e-mail address to send you the full wireshark report and the telnet server were are using right now.

-----------------------------------------------------------------------------------------------------------------

Please pay attention to the fact that for some reason(time_out?????) one client stops asking for conection to the telnet host, after that the only way to revive it is rebooting.

----------------------------------------------------------------------------

Only S2E is on and Telnet Server(192.168.1.103) is Down

The S2E request 2 id's (port 0 and port 1) to Telnet Server

13 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4097)

14 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

15 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

16 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4097)

24 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

25 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4097)

48 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

49 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4097)

60 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

61 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4097)

65 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

66 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4097)

71 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

72 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4097)

74 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4099)

After this point S2E requests only one ID!!!

75 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4100)

85 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4100)

88 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4100)

xxx  Telnet Server (192.168.1.103) is UP

93 - Conectado no Cliente 1 (192.168.1.240 - id:4100) - The ID: 4100 is accept

     But only one ID S2E (192.168.1.240 - ID:4100) the other port is not connected

.

.

xxx - Telnet Server is down

159 - Desconectado Cliente 1 (192.168.1.240 - id:4100) - The ID:4100 is disconnect

.

182 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4105)

191 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4105)

206 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4105)

210 S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4105)

xxx  Telnet Server (192.168.1.103) is UP

Telnet Server connect only one ID:4105, the other port not.

.

.

xxx - Telnet Server is down

232 - The S2E is disconnect, iD:4105 is off.

262 - S2E is turn off

xxx - S2E is turn on

294 - S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4097)

295 - S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

296 - S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

297 - S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4097)

314 - S2E (192.168.1.240) -> Telnet Server (192.168.1.103) - PORT (4098)

xxx  Telnet Server (192.168.1.103) is UP

In this case is connect Port 0 (Cliente 1) and Port 1 (Cliente 2)

323 - Conectado no Cliente 2 (192.168.1.240 - id:4097) - The ID: 4097 is accept

325 - Conectado no Cliente 1 (192.168.1.240 - id:4098) - The ID: 4098 is accept

.

.

xxx - Telnet Server is Down

809 - Desconectado Cliente 1 (192.168.1.240 - id:4098)

817 - Desconectado Cliente 2 (192.168.1.240 - id:4097)

Thanks for the trace. I'm pretty sure we are seeing different problems so I'll concentrate on trying to reproduce what you are seeing here. Your server is obviously behaving differently from the one I'm testing with since I see the local port number increment on every request whereas your trace seems to indicate that the same port numbers are reused (presumably indicating that the server has rejected the connection attempt before the next retry). In my scenario, I am seeing the S2E code attempt to reconnect approximately every 2mS which strikes me as rather too frequently!

It would still help me to know what your S2E port configuration settings are so that I can duplicate them (as far as possible). Also, can you provide any information on the actual server you are talking to? Is it using software that I could install here rather than using Windows server in case this is important? If you want to send me the full Wireshark trace (not that I reckon I will be able to tell much from it in isolation), feel free to send a private message and I'll get back to you with an email address.

  When the S2E keep trying to connect with the same port number the telnet server is down. so it is independent of server behavior. we placed in the trace becouse we consider that the S2E client is dying due to the lack of server response for a period of time.

  Our two S2E clients are configured as clients (local IP:192.168.1.240)  (server IP:192.168.1.103)  (remote server port: 250)   (both serials 9600,8,n,1) (protocol: Telnet even though we do not know what raw really means!)

  The server we are using is a INDY server, programmed in Delphi, however since we can see the failure when the server is down, I guess it is not important.

  We made the following test:  Do not start any server, just start the S2E and observe that one client stop comunicating. in our test the ports ranged from 4097 to 4100 when reache 4100 only one client is making requests it took 15segs most attempts.

I'm still not having much luck reproducing this. With the Windows telnet service stopped, however, Windows TCP/IP stack resets each connection attempt so the board is getting a response to each connection request. I'll try setting up a bogus server IP address and see if I see the behavior you mention.

Please could you confirm the version number of the S2E software you are running and, just to make sure we are on the same page, update to the very latest release if you have not done so already?

In "raw" mode, no data on the connection is parsed by the S2E software prior to passing it on to the serial client. In "telnet" mode, any telnet control sequences (to set serial port parameters, for example) are parsed out of the stream and applied by the S2E then removed from the stream and not forwarded to the serial client.

Ciro,

I'm now able to reproduce this problem - thanks for your information. This effectively restates what you have already said but it helps me keep track so here's what I did.

1. Set up 2 S2E modules on the same LAN. Configure 1 with both ports set as servers on port 23 and 26 respectively. Configure the other with both ports as clients one connecting to port 23 and the other to port 26 on the server module.
2. Leave the server powered down.
3. Power up the client with WireShark grabbing everything.
4. After a minute or so, power up the server.
5. At this point, the Port 0 on the client successfully connects to the server but no connection attempt is seen from Port 1 to the server. Looking at the session state, several connection attempts were made before the server came online but the Wireshark trace shows no attempt to initiate a telnet connection on the TCP port configured for client Port 1.

The code to handle the connections is completely symmetric as far as I can tell so I suspect this is likely some race condition or interaction between the two connection attempts for each of the client ports. If I reconfigure the "client" Port 0 as a server, Port 1 connects successfully using the same scenario described above.

Investigation continues...

Dave,

Thanks for the effort, but in fact in our scenario both clients point to the same server in the same port. I don't know if it could make things more complex, but that is the way we plan to use it.

by the way, you said that in telnet mode S2E could extract config. information from the stream before sending it to the serial port, I couldn't find these "telnet control sequences" format especialy regarding configuring the serial ports. I am also want  to be able to configure S2E via stream, not only via HTTP.