Orphan Logins - Netware

This is a discussion on Orphan Logins - Netware ; Hi- We're on a centrally managed Netware 5 network, using client v. 4.83 sp1. About 10 of the 40 clients I support in our particular office are experiencing login problems, indicating that they are trying to log in to too ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: Orphan Logins

  1. Orphan Logins

    Hi-

    We're on a centrally managed Netware 5 network, using client v. 4.83 sp1.
    About 10 of the 40 clients I support in our particular office are
    experiencing login problems, indicating that they are trying to log in to
    too many machines simultaneously (by default, our users get 4 concurrent
    logins), although they aren't logged in anywhere else. Nlist shows multiple
    logins, all from the same node.

    Our admins insist that these users are not properly logging off their
    computers, thus leaving orphan logins. Their solution has been to bump up
    the number of concurrent logins they're allowed. One user currently shows
    10 logins from the same node.

    This has been an ongoing issue for months now (our users generally log off
    and shut down their computers overnight, on weekends, and over vacations),
    and has led to much hate and discontent, focused primarily at me.

    So, I have a couple of questions:

    -Shouldn't the server clean up any inactive logins after a certain length of
    time?

    -Is there some account setting - or anything else - that would cause this
    to happen on certain account, but not others (I find it hard to believe that
    only 25% of our users would manage to screw something up; if something can
    be screwed up, we're generally in the 90 - 95% range).

    Any insights would be greatly appreciated.

    Patrick Sauter



  2. Re: Orphan Logins

    On Mon, 13 Dec 2004 10:45:00 -0500, pjs wrote:
    > We're on a centrally managed Netware 5 network, using client v. 4.83 sp1.
    > About 10 of the 40 clients I support in our particular office are
    > experiencing login problems, indicating that they are trying to log in to
    > too many machines simultaneously (by default, our users get 4 concurrent
    > logins), although they aren't logged in anywhere else. Nlist shows multiple
    > logins, all from the same node.


    This is a long standing problem, but it is actually multiple problems with
    the same symptom. What you're seeing is the "stuck network address"
    attribute. If you search via Google's "advanced group search" feature,
    you'll find *lots* of references to this. I won't repeat all I've written
    before about this one.


    > Our admins insist that these users are not properly logging off their
    > computers, thus leaving orphan logins.


    Somewhat harsh criticism, but your central admins haven't done their
    homework on this one. They're wrong, plain and simple.


    > Their solution has been to bump up
    > the number of concurrent logins they're allowed.


    They also haven't researched their solution very well. There are at least
    three publicly available ways to remove stuck Network Address attribute
    values. Depending on your rights to the User objects affected, you may be
    able to do this (search on "remadr.exe" on http://support.novell.com), or
    it may be something your central admins would have to do.


    > This has been an ongoing issue for months now (our users generally log off
    > and shut down their computers overnight, on weekends, and over vacations),
    > and has led to much hate and discontent, focused primarily at me.


    Sorry for that, but I understand what you mean here. I've been fighting
    this problem since our NW4.10 in 1996. I had it with NDS v6 and v7. I had
    it resolved with NDS v7, then we upgraded to eDir 8.5 and it came back.
    I've been fighting it ever since.

    There are multiple bugs, in NetWare, eDir, and Client32, all of which have
    this as a symptom. You say you're using Client32 4.83, so that's one
    upgrade (v4.9 SPsomething is out) you should definitly explore. On the back
    end, being up to date on the OS version (NW 6.5 is current, you're using
    v5.something), Service Packs (see http://support.novell.com) and version of
    eDirectory (eDir 8.7.3.3 is current) will definitly help make the problem
    happen less frequently. Here, with about 30K users on the network, and
    dealing with tens of thousands of login/logout events per day, I'm down to
    seeing this problem happen only once every couple of weeks.


    > -Shouldn't the server clean up any inactive logins after a certain length of
    > time?


    Yes. Google for "watchdog" and "monitor connection" and "stuck network
    address", especially in the novell.support.* newsgroups. You should also be
    able to find "watchdog" in the online documentation.


    > -Is there some account setting - or anything else - that would cause this
    > to happen on certain account, but not others


    Nope. Well, unless you count not using the maximum allowed concurrent
    connections feature a setting. You could turn that off, and would then not
    have this problem.


    > (I find it hard to believe that
    > only 25% of our users would manage to screw something up; if something can
    > be screwed up, we're generally in the 90 - 95% range).


    It's not the users. It's not you. It's the software.


    > Any insights would be greatly appreciated.


    Not to sound pompous about it, but I can truthfully say that I know way
    more about the Network Address attribute than I ever wanted to, and likely
    more than anybody else you're likely to find on-line or off. You might want
    to add my name ("gersic") to your Google usenet archive queries, where
    you'll find lots under "stuck network address". Happy (?) reading.


    --
    | David Gersic dgersic_@_niu.edu |
    | I love Christ and hope to meet a Christian. - Ghandi |
    | Email address is munged to avoid spammers. Remove the underscores. |

+ Reply to Thread