[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference noted::pwv50ift

Title:Kit: Note 4229; Please use NOTED::PWDOSWIN5 for V4.x server
Notice:Kit: Note 4229; Please use NOTED::PWDOSWIN5 for V4.x server
Moderator:CPEEDY::KENNEDY
Created:Fri Dec 18 1992
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:4319
Total number of notes:18478

4198.0. "V5.0E - loosing access to a network drive, user mapped to GUEST" by LNZALI::BACHNER (Mouse not found. Click OK to continue) Mon Mar 10 1997 10:48

Since upgrading from V5.0D ECO2 (?) to V5.0E on OpenVMS VAX V7.0, I frequently
run into the following problem:

I can connect to a network drive without problems. The LANMAN user is mapped to
a specific OpenVMS account. I can work with the share, read, modify, add files.
All of a sudden, I get the message that I'm not allowed to access the files on
this drive. $ NET SESSION on the PWRK server shows that there a two sessions
from my workstation (running Win NT/Alpha 4.0 SP2), one using the correct user
name, the other one using GUEST.

Am I doing anything wrong ?  Do I need to change some configuration data ? 
So far, I have been unable to find a consistent, reproducible way to make this
failure appear - but it will appear for sure, sometimes three times a day,
sometimes once in two days. Disconnecting and reconnecting the share fixes the
problem in most cases - somtimes I need to $ NET SESSION \\my_wkst /DELETE to
make it work again.

Thanks for your ideas,
Hans.
T.RTitleUserPersonal
Name
DateLines
4198.1I have the same problem.DUNKLE::MCDERMOTTChris McDermott - Software Janitorial ServicesMon Mar 10 1997 13:478
I have the same problem from an NT 4.0 client.  The client makes two connections
to the server.  Once as the user I am logged in as and once as guest.  The file 
server is always using the first session no matter what session the SMB comes in
over.  

I thought that I file a QAR on this but I can't find it so I'll enter another.

Chris
4198.2re: -.1 QAR 402 in the PW-OVMS-V50D-ECO QAR databaseDUNKLE::MCDERMOTTChris McDermott - Software Janitorial ServicesMon Mar 10 1997 14:044
Note that this *may* be fixed in V50E-ECO1.  I am going to test it and I will
report back.

Chris
4198.3Not fixed yetDUNKLE::MCDERMOTTChris McDermott - Software Janitorial ServicesMon Mar 10 1997 15:345
This is not yet fixed in V50E-ECO1.  I am not in a position to say whether or
not it will be fixed before ECO1 ships.  As noted previously, the problem has
been QARed and the project leader is aware of it.

Chris
4198.4I am in a position...PATRLR::MCCUSKERMon Mar 10 1997 16:544
It will not be fixed in ECO1...  ECO1 is in the very final moments of
qualification.

It will be addressed quite seriously for V50F.
4198.5V5.0F - this would mean another 3+ months ?LNZALI::BACHNERMouse not found. Click OK to continueTue Mar 11 1997 05:516
> It will be addressed quite seriously for V50F.

Any chance for an intermediate patch ?  This has become a real nuisance.

Thanks,
Hans.
4198.6Downgrade to V5.0D-ECO3 will do it!STKAI1::BLUNDBERGCollission speed full!!!Tue Mar 11 1997 08:074
    Go back to V5.0D-ECO3, this worked around the problem for me.
    
    Bjorn
    CSC Sweden
4198.7PATRLR::MCCUSKERTue Mar 11 1997 10:2913
>.5 Title:  V5.0F - this would mean another 3+ months ?

I think V50F will be on a little shorter cycle than usual, but also, as
always, no commitments in this forum.

.5> Any chance for an intermediate patch ?  This has become a real nuisance.

Sure, IPMT it first.

.6> Go back to V5.0D-ECO3, this worked around the problem for me.

Thats news, I didn't realize we broke it in V50E.  Bjorn, did you IPMT it and
I just didn't realize the case was already in engineering?
4198.8only a WinNT issue ?LNZALI::BACHNERMouse not found. Click OK to continueTue Mar 11 1997 13:196
Is this a problem only for WinNT V4.0 workstations or for Win95/WfW 3.11 as
well? I need to know before I recommend the customer mentioned in the next topic
which version he should upgrade to.

Thanks for all your help,
Hans.
4198.9CFS.49378STKAI1::BLUNDBERGCollission speed full!!!Wed Mar 12 1997 03:078
    I've issued an IPMT on this problem, I've also recieved a set of
    new images that "might solve it", unfortunatelly I've been unable
    to reach the customer so I don't have any feedback on that yet.
    
    One thing is clear though...I've seen three different cases where
    a downgrade to V5.0D-ECO3 has worked around the problem.
    
    Bjorn
4198.10NT only ?STKAI1::BLUNDBERGCollission speed full!!!Wed Mar 12 1997 03:118
    Hans,
    
    I,ve only seen the problem occur when NT workstations connect to the
    V5.0E server.
    I believe it has to do with the server authenticating the client on
    a session without username.
    
    Bjorn
4198.11UTRTSC::SWEEPI want a lolly...Wed Mar 12 1997 03:5813
    We noticed that you get this extra "session" when you have
    a my-computer screen active on NT and using a logon name
    which is != to the connection name.
    
    Can you check the UAS records of both users, are there any
    logon hour restrictions.
    
    It looks like the servers gets some uid context mixed-up.
    We are right now testing a large nr of multi-session fixes,
    but I don't think this is 1 of the fixes. We'll simulate
    it and report back.
    
    Adrie
4198.12may I try it ?LNZALI::BACHNERMouse not found. Click OK to continueFri Mar 14 1997 05:4239
Adrie and Bjorn,

thanks for all the details.

I'd be happy to test the new images mentioned in .9 on our internal system
(where I observe the problem) if someone could provide a pointer to them. We run
the PATHWORKS software on a VAX.

The hint with using a share logon name different from the workstation logon name
might point into the right direction.

Here's some history on my problem - I hope to recollect it correctly:

My workstation belongs to the same domain as the PATHWORKS server, and I always
log on to the domain. At some point in time (I think we still had the V5.0D ECO
something software running, the VAX was a member server) I could not (no longer)
connect to my shares properly. I had only read access to my shares though I was
their owner.

We tried some things like making the server a standalone server and then a
member server again but I was unable to use the shares properly.

Then I created a new account by copying the old one and just gave it a different
name (using the account manager for domains). I specified this new account in
the 'connect to share' dialog box and it worked fine.

Some time later, probably after upgrading to V5.0E, I experienced the problem
described in .0. While experimenting to get rid of the problem, I found out that
I could use my 'standard' account again - but with the same problem as the
alternate account.

Is it possible that PATHWORKS is confused by the fact that my domain account has
the same name as my workstation ?  More than that, the account exists in the
workstations local account database as well, but that should not matter when I
log on to the domain.

Hope this helps you to track the problem down.

Hans.
4198.13Hope it will be solved some day.STKAI1::BLUNDBERGCollission speed full!!!Thu Mar 20 1997 04:5922
    The customer that first encountered the problem didn't want to install
    the new images for V5.0E, he didn't want to risk ending up in a situation 
    where he had to perform a downgrade to V5.0D-ECO3 again.
    
    I have now however a second customer that has encountered the same
    problem and he was happy to try the images, unfortunatelly they didn't
    solve the problem though.
    
    This customer has more than a dozen pathworks servers and after they
    started to upgrade their workstations to NT4 they also encountered 
    this problem.
    This customer has downgraded one of the servers to V5.0D-ECO3 just to
    see if the problem disappeared and as one might have suspected it did!
    
    The story with CFS.49378 goes on... unfortunatelly I'm not sure I'll see 
    the end of it since I'm leaving Digital at the end of next week.
    Speaking about that I'd like to thank all of you that has answered
    questions and helped me out in this conference over the past two years.
    
    Thanks
    
    Bjorn
4198.14Fixed in ECO1CPEEDY::FLEURYThu Mar 20 1997 08:345
    RE: .-1
    
    This problem is resolved in V50E-ECO1.
    
    Dan
4198.15UTRTSC::SWEEPI want a lolly...Mon Mar 24 1997 06:0423
    We made a new fix for e eco 1 or e ssb via ipmt channels
    available.
    
    The story:
    
    NT sends a sesssetup smb with a NULL username. This means
    that when you do a net sessions you see a valid username
    and a null username.
    
    Unfortunately there was a slight problem that when the
    NULL session request came along, and the active session
    was the origional username session that the active user
    was marked as a NULL session. This meant that this user
    could get access problems on his already connected shares.
    
    So a possible bug in NT (the empty username session) caused
    a bug in PW. The fix now makes sure that we mark the
    correct session as a NULL session (= the new session).
    
    This fixes the access errors but still shows an empty
    username session (which is valid because NT sends us one).
    
    Adrie
4198.16still experiencing the problemTLE::MICHAUDLisa Michaud, DTN 381-0879Mon May 05 1997 16:5620
    I was seeing the same problems and upgraded to ECO 1, but the problems
    are still there.
    
    It's random- a user will have a session established, then it will
    disappear and when he tries to reconnect, he gets "network access is
    denied".  Sometimes users can connect to the domain alias instead
    (there's only one node at this time, but it has an alias), or
    visa-versa if the alias connection has died.
    
    If I go into Admin and look at session status, some of the users do
    have 2 connections, one under their own name and one under GUEST
    (although if you zoom into that connection, at the top it says
    "Username" and it gives their domain user names).
    
    So, did ECO 1 fix the problem, or not?  I think I'm going to go back to
    V5.0D...
    
    Thanks-
    
    Lisa
4198.17Right, ECO1 has same problemVMSNET::P_NUNEZMon May 05 1997 20:305
    Lisa,
    
    The fix isn't in v5.0E ECO1 either - it too needs a patch...
    
    Paul
4198.18can I get the patch?TLE::MICHAUDLisa Michaud, DTN 381-0879Tue May 06 1997 09:005
    So can I get this patch, or do I have to go back to 5.0D?
    
    Thanks-
    
    Lisa
4198.195.0D?TLE::MICHAUDLisa Michaud, DTN 381-0879Tue May 06 1997 14:144
    OK, so I'm not getting an answer on the patch.  Can someone point
    me to a PATHWORKS V5.0D kit?
    
    
4198.20Submit IPMT = acquire patchCPEEDY::FLEURYTue May 06 1997 16:095
    As stated elsewhere in this conference: to get a patch, submit an IPMT
    case.  The patch is available for V50E-ECO1.
    
    Dan
    
4198.21thanks...TLE::MICHAUDLisa Michaud, DTN 381-0879Wed May 07 1997 09:296
    Sorry to have wasted anyone's time.  I was confused by replies 14
    and 15, which stated that the fix was included in ECO 1.
    
    Thank you for the information you have provided.
    
    Lisa
4198.22IPMT for internal problem?TLE::MICHAUDLisa Michaud, DTN 381-0879Wed May 07 1997 13:0911
    It's my manager's understanding that IPMTs are for customer problems.
    
    My problem is on an internal system.  Does this mean that I still have
    to go through the IPMT process to get the patch?  
    
    No one here can even help me do that, because they've never had to do
    this before to obtain a patch.
    
    Thanks for your help-
    
    Lisa
4198.23VMSNET::mickey.alf.dec.com::s_voreSmile, Mickey's watching! [email protected]Wed May 07 1997 16:485
talk with the IPMT Problem Managers - they'll tell you that you can 
use the IPMT process to get any problem fixed - not just a 
customer's.