| T.R | Title | User | Personal Name
 | Date | Lines | 
|---|
| 527.1 | FT10D.05 has some RADIUS changes in it | IROCZ::RRICHARD |  | Thu Mar 20 1997 13:06 | 19 | 
|  | 
  FT10D.05 contains two changes having to with the way RADIUS reply messages
  are handled.  Is he using RADIUS?  I reviewed the changes and didn't see 
  anything that would account for the problem.  Has the customer gone back and tried BL10D
  to verify that it's the software?
  What client is he using on his Windows 3.1 PCs?
  Are the Win 3.1 PCs using the same type of authentication?
  He might be able to isolate the problem a little more by zeroing, then
  monitoring the port's security counters while logging in.   This would
  at least verify that authentication requests are being sent to the host.
  Has he tried bringing up the connection manually to see if he can identify
  what part of the process is failing?
 | 
| 527.2 | more info... | CSC32::D_SHAVEY |  | Thu Mar 20 1997 15:43 | 15 | 
|  |     More information:
    
    Nothing has changed on these Windows 3.1 PC's, and he tested it and
    they can authenticate on the BL10D-40 image fine. They have Netscape
    3.0 and are using the shiva dialer which is part of the personal
    edition Netscape 3.0 kit. The authentication method here is PAP. His 
    DRAS server is an Windows/NT box. He says with this FT10D.05-40 patch
    with authentication turned off, the windows 3.1 pc's can dial in fine 
    with a PPP connection. He will look at the security counters after
    trying to dial in with the windows 3.1 pc's, but right now the total 
    packets send and valid packets received match.
    
    Darrell Shavey
    CSC32::D_SHAVEY
    DTN 592-4712
 | 
| 527.3 | any ideas? | CSC32::D_SHAVEY |  | Mon Mar 24 1997 13:12 | 9 | 
|  |     Any ideas here. I can't have him go back to the previous patch, because
    of the crashing problem. But with authentication turned off, these
    windows 3.1 pc's can dial into the decserver fine.
    
    thanks,
    
    Darrell Shavey
    CSC32::D_SHAVEY
    DTN 592-4712
 | 
| 527.4 | we need more information | IROCZ::RRICHARD |  | Tue Mar 25 1997 09:46 | 33 | 
|  | 
  The RADIUS related changes made in FT10D.05 only affect post authentication
  processing.  Specifically, the first change does some cleanup after failed
  authentication requests, and the second change inhibts the output of the text
  contained in the reply messages if PPP is active.  Since the DRAS accounting
  log doesn't contain any accounting records for the failed authentication it
  would appear that the authentication is not progressing far enough to
  encounter the changes.  I don't have the Shiva client available so I tested
  with Trumpet and Windows for Workgroups V3.1.  I was not able to reproduce
  the problem.  That could be because I didn't have the customer's DRAS
  user profiles or DECserver configuration.
  What are the differences between the DRAS user profiles for the Windows 95
  and the Windows 3.1 users?  Can we get a sample of both?
  Has the customer monitored the port security counters during authentication
  to determine if the port is at least attempting to authenticate?
  Are there  any differences between the port configurations for Windows 95
  and Windows 3.1 users?  Can we get a copy of the LCP, IPCP, and PORT
  settings?
  Does the Shiva client work with a script?  If so can we get a copy of it?
  Is there a way to manually log in then start PPP with the Shiva client?
  Windows 95 allows the user to bring up a window, then log in and start PPP
  does Shiva support this?  If so can it be used to determine if any error
  messages are being printed during authentication?
  Does Shiva produce a detailed log like Trumpet does?  If so can we a trace
  of both types of login?
  
 | 
| 527.5 | more info....hope to have more | CSC32::D_SHAVEY |  | Tue Mar 25 1997 15:30 | 14 | 
|  |     The windows 3.1 pc's are using the Netscape Personal Edition and the
    dial in script similar to dial up networking is in here. He said that
    Netscape just calls it the shiva script. These windows 3.1 pc's dial
    into the same decserver 90m's as the windows 95 pc's for
    authentication, and if the authentication characteristic is turned off
    on the DECserver, the windows 3.1 pc's dial in fine and can make their
    PPP connection, and all works fine. He will be sending me the LCP, IPCP
    and port settings and I will post them here. So the windows 3.1 pc's
    make the ppp connection fine, just can't be authenticated. He has no
    tools to get a trace, so as a last resort I can get field service
    involved to get one if needed.
    
    Darrell
           
 | 
| 527.6 | We could set something up in LKG for testing | IROCZ::RRICHARD |  | Wed Mar 26 1997 07:43 | 22 | 
|  |   
  Thanks for the additional information.  The best thing to do at this point
  may be to set something up in LKG and ask the customer to dial in so we
  can monitor the authentication attempt.  Let me know if you want to take
  this route.  It will take about a day to get everything setup.  In order to
  emulate the customer's configuration I'll need a sample of a DRAS user 
  profile in addition to the port, LCP, and IPCP information he's collecting
  for us.
  What are the differences between the DRAS user profiles for the Windows 95
  and the Windows 3.1 users?  Can we get a sample of both?
  Has the customer monitored the port security counters during authentication
  to determine if the port is at least attempting to authenticate?
  Can the customer provide a copy of the dialin script his customers are using?
  Does Netscape produce a detailed log like Trumpet does?  If so can we a trace
  of both types of login?
  Would it help if I called the customer?  If so mail me the information at
  twoby4::rrichard.
 |