Conference iosg::all-in-1_v30

    Hoi Luc,
    
    I don't know if this is exactly your problem but there was a STARS
    article about it. See below.
    
    Groetjes, Sjaak Bosman
    
    ---------------------------------------------------------------------
    
Scan For Free Time Fails For Remote Users: 'Network Not Available'


COPYRIGHT (c) 1988, 1989, 1990 by Digital Equipment Corporation.
ALL RIGHTS RESERVED. No distribution except as provided under contract.


PRODUCT: ALL-IN-1 V2.4

SOURCE: Customer Support Center/Atlanta USA

PROBLEM:

Since upgrading the system from ALL-IN-1 V2.3 to V2.4 the following problem
has been identified when attempting to scan for free time of remote
users.

In V2.4, attempts to scan a remote user calendar from free time using
Time Management (TM EV SC) fails indicating 'Network not available' 
as below.  This problem occurs when the remote username contains a
period (.).  Scanning local users appears to work correctly.


                       Available Schedules for 30-Oct-1990


                          8   9   10  11  12  1   2   3   4   5

COMICS::TEST . JO           ( Network not available )

CAUSE:

The problem was due to the code incorrectly parsing the username which
was then unrecognizable on the remote node.  The Scan code would then
return the error status "Network not Available"


SOLUTION: 

The problem you describe is a restriction in V2.4 of ALL-IN-1.
Your observations will be assessed for inclusion in a future
release of ALL-IN-1.

This problem is rectified in Patch K604.  To resolve this problem it
is recommended that you update your system to the current ALL-IN-1 
V2.4 patch level by installing up to the latest available patch.

Refer to the related articles for details:

 ALL-IN-1 V2.4 UPDATE V2.0 Rev 1.1 (K603) Patch Kit Information
 Current ALL-IN-1 V2.4 Patch Kit Information

HOW TO OBTAIN THE LATEST CORRECTED SOFTWARE PATCH FOR V2.4 SYSTEMS:


This software correction is available through the DSNlink VTX Patch
Application.  For more information on how to use this application,
please refer to another article (in the Text Search Database called
DSNLINK) entitled: 

     How To Downline Load Patches From DSNlink VTX

While in the DSNlink VTX Patch Application, please use the keyword
"ALL-IN-1" to find the patch titled: "ALL-IN-1 V2.4 Patch Kit".

If you do not currently have the DSNlink for VMS product installed on
your system, or have questions on DSNlink, please contact our DSNlink
Hubroom at 1-800-332-8014. 

    Hello Sjaak ,
    
    This is not the same problem . I saw this article also.
    
    Thanks for your reply.
    
    Regards.

        What's the address you are entering at the attendee line?
        
        Dave Zaniewski

	There was another bug in the Scan code that caused the "Network not
available" message to be generated at unusually frequent intervals.  The patch
was created some time late last year, and was passed on to IOSG::  As far as I
know, it has been incorporated into the V3.0 code.

    Message router isn't involved in free time scans. The code connecys
    directly to a server running on the node which holds the calendar
    you're scanning.
    
    The free time scan code parses the address typed to find the actual
    nodename to which to connect to do the scan. This parse can get
    confused in some circumstances.
    
    The fact that you're using a remote message router might be a clue. As
    Dave asked in .3, what address is the user typing?
    
    Dave�
    
    

    Hello,
    
    There is a little typing error in .0
    
    System B is running ALL-IN-1 and uses Message router from System A
    
    The addr. is USERNAME@REMOTE_MAILBOX_system_B@LOCAL_ALLIN1_mailbox
    
    Regards,
    
    Luc.

    Then I suspect you've found another of the problems with Free Time Scan
    (although we'll need to wait for a TM developer - a what? :-) - to
    answer to be sure).
    
    As I said, TM has to parse the address typed by the user to find out
    the node on which the TM server (and the user's calendar) lives.
    
    Normally, it can do this - if the address is like SHOVE @A1 @TRON, it
    assumes that the last address term is the nodename, so it connects to
    the TM server process on TRON:: - which is correct.
    
    However, there are (at least) two cases where this won't work:
    
    1. (mostly applies in Digital) - area routing - given an address like
    Dave Shove @REO - how does it work out that the node it needs to
    connect to is TRON:: ?
    
    2. Remote message routers - given an address like User @RemMailbox
    @MrSystem - it will connect to the server on MrSystem:: (the system
    which runs the remote Message Router). This isn't the system that
    contains User's calendar files.
    
    There isn't any solution to these problems, at the moment. Indeed,
    we're currently puzzling over how to fix them for Possible Future Time
    Management Products (PFTMPs ?). The only way seems to be to hold the
    actual server-node for each user in some kind of directory (such as
    DDS) - this is obviously far too big a change to put into ALL-IN-1 IOS
    (especially as IOS doesn't necessarily use DDS).
    
    Sorry,
    Dave.

We have the same problem, however we have a consistent mailbox naming.

i.e. 	carlier @ a1_nodeA @ mr_node

where

	ALL-IN-1 account name is 'carlier'

	MR mailbox for 'nodeA' is 'a1_nodeA'

	One MR on network on 'mr_node'

So my question is (before I spend a lot of time coding it) can I change 
some scripts/forms to fool the TM scan routine to use CARLIER@NODEA in 
place of the above ?

If so which/what ?

Of course the eventual mail will have to use the full 
	carlier @ a1_nodea @ mr_node.

Other ideas are also welcome ?

MR on every node is NOT one of them at this time.

Regards

John.

        Hi John,
        
        What you suggest should be possible.  The script TMMESTRIP already
        ignores the stuff between the @'s already.
        
        A little end user re-education might be more useful.  Both
        username@A1@nodename and username@nodename work right now.
        
        Dave Zaniewski

re: .-1

Dave,


Thanks for the reply but I think you miss the point. 

The addressee is on, in my example, NODEA, and not on MR_NODE.

I know the SCAN routine will work with 'user@A1@NODE' and 'user@NODE', but 
with user@node the eventual mail message to plan the meeting will fail.

In my customers situation it is as follows

         ---------    Meeting          -------
         |MY_NODE|<-- planner          |NODEA|<-- Here is 'USER' account
         ---------                     -------     with TMSV + calendar



                        --------- 
                        |MR_NODE|<-- Here is MR
                        --------- 

On MY_NODE user has a Mail address of USER@A1_NODEA@MR_NODE and therefore 
the SCAN option looks at MR_NODE and not at NODEA.

If I give an address of USER@NODEA then the scan works but the mail to plan 
the eventual meeting then fails....

Regards

John.

Finally got on site by the customer yesterday and managed to fix the 
situation for them.

What I did was to add code in the TMSCAN.SCP just before and just after the 
call to CAL SCAN.

This before code examined the entries in CAL$SCROLL_ATTENDEES and if found 
to be be a remote entry modified the address to include the node of the 
attendee, based on the mailbox.

i.e.	JONES @ A1_xyz @ MR_NODE is modified to be

	JONES @ A1_xyz @ xyz @ MR_NODE.

This address is OK for the CAL SCAN.

After the SCAN I then remove the '@ xyz' or I get problems matching the 
meeting replies with the meeting attendee, as the reply comes from 
JONES @ A1_xyz @ MR_NODE.

I tried to modify the CAL$SCROLL_ATTENDEES in a FOR loop but this did not 
work. You must set up an index as follows

	get #number = 0
.label loop

	compute #number = #number + 1
	get #text = cal$scroll_attendees.text[#number]

	! process #text

	write change cal$scroll_Attendees number = #number , -
				 	  text = #new_text
	goto loop

BEWARE this all changes in V3.0.

Regards

John.