[Dxspider-support] some sort of new lockout problem

[Charlie Carroll]

Dirk:
The node is being restarted so that I regain connections to other nodes and
so that old users/nodes are treated as old users/nodes and not new users.

I probably am using some combination of shutdown and Control C.  The node is
not crashing.

I tried using the export_user command and was expecting to see a file named
user_asc with a timestamp of today.  That is not happening.  My user_asc
file is still dated 5/26/01.  However, my users.V2 file is dated today.

Even after a re-start, things are not "as they were" in the recent past.
Sorry I can't be more definitive.  I was supposed to be home tonight, but am
stuck at a customer site for another day.  I just don't have the ability to
really see real well what's happening at home.  More after I get home
tomorrow.

charlie, k1xx

----- Original Message -----
From: "Dirk Koopman" <djk at tobit.co.uk>
To: <dxspider-support at dxcluster.org>
Sent: Wednesday, May 30, 2001 11:50 AM
Subject: Re: [Dxspider-support] some sort of new lockout problem


> On 30 May 2001 17:31:04 +0200, Aurelio -PA3EZL- wrote:
> > Hello ,
> >
> > I'm sending this mail on behalf of Charlie-K1XX who cannot send e-mail
in
> > this moment
> >
> > ===
> > A new problem has now shown its ugly face.  Some short period of time
after
> > a restart, the W1HR node "forgets" who users are!  If I attempt to login
as
> > k1xx-6, the node greets me as if I were an entirely new user.  It also
> > happens when other nodes attempt to connect to me.  They also are
greeted as
> > a new user.  Plus, my outgoing connects fail.
> > If I restart the s/w, all is back to the original cond
> > Over some period of time (seemingly hours at most), the condition
reoccurs.
> > Unfortunately, I am traveling so it's difficult to do a lot of
> > troubleshooting.  There is an indication that the problem may be
associated
> > with the new lockout code.
>
> Hum...
>
> How/why is the node being restarted?
>
> Specifically are you shutting it down using the 'shutdown' cluster
> command or is it crashing or having ^C done on it?
>
> It may worthwhile recreating the user file from the ascii version which
> is produced every weekend. You can also produce a new one on demand by
> using the 'export_user' command.
>
> To recreate the userfile:-
>
> 1. shut the node down.
> 2. cd c:\spider\data (/spider/data)
> 3. perl user_asc
> 4. restart the node.
>
> There is a problem with DB_File (the 'binary' form of the user file)
> which means that it must be closed cleanly in order for it to work
> correctly - I go to a _GREAT_ deal of trouble to make sure this happens.
>
> Also on older versions of DB_File, it was too easy to create duplicate
> entries.
>
> Either of these problems may exist on the Windows versions. The only
> cure is to edit the 'user_asc' file, removing duplicate entries that
> aren't what you want and then doing the user file recreate procedure
> above.
>
> I haven't seen this problem however for at least a year and AFAIK nobody
> else is experiencing it (says he putting tin hat firmly on head).
>
> Dirk
>
>
> _______________________________________________
> Dxspider-support mailing list
> Dxspider-support at dxcluster.org
> http://www.tobit.co.uk/mailman/listinfo/dxspider-support
>