Conference 44.370::system_management

    1) RDGE00 had a new network database installed over the weekend
    This is somewhat rather larger than the old one and consequently
    requires more system resources. 
    2) RDGE28 cannot have a large DECNET database until it can get rid
    of the PDP DMR/DMC lines. A large database has been tried before,
    it simply makes RDGE28 crash! 
    
    1 is being fixed by tuning and rebooting
    2 has a fix on the way. The DMR/DMC boards will be moved to a dedicated
    11/750 Vax which will handle DECNET traffic only. This should allow
    U28 to run a full network database. I expect the 11/750 to be installed
    around about 30th May.
    
    Jerry Thompson

    OK Jerry, thanks for the info. 
    
    Next question; why do you load the network databse into RDGE00
    when if you type 'SHOW NET' it tells you that it is a 
    non-routing node that does not have any network information
    and that the router is RDGMRC.
    
    Peter	(just_a_PDP_man_trying_to_make_a_living)
    

    Hi Peter,
    
    Re .-1
    
    The node database on RDGE00 specifies the nodes that RDGE00 knows
    about. The fact that RDGE00 is a non-routing node, ie traffic can't
    pass through it, is a sepperate issue.
    
    If RDGE00 had no node database it wouldn't be able to communicate
    with any other node (by specifying a node name rather than a network
    address).
    
    Hope that this makes things as clear as mud
    
    Chris 

    Ah ha....
    
    does this mean that even when RDGE00 has no 'node database", that
    I can still do a set host if I know the node number...
    
    If this is the case, where can I get  a list of them, and all
    my problems will be solved (well, some of them.)
    
    Peter.
    

	Peter,

	Yes, you can access any node in the world from a machine with an
	empty node database if you only know the address of the desired node.
	
	Dead useful that.

	There are at least two ways to find out the address if you know the 
	name:
	
	  a.  Use VTX's GENEVA database, option 12 on main menu. This lists
              all the known names and their addresses.

	  b.  $MC NCP TELL RDGMRC SHOW NODE name
              This will use the database on RDGMRC (a pretty good one) to get
              the address if it knows the name.

	The address will be in the format:

		n1.n2

	If n1 is the same as n1 on your current node, then you need only 
	specify n2 as the address.

	If n1 is different, then you must do the following calculation:

		address = (n1 * 1024) + n2

	Hope this helps,

	Howard

    Great, Howard - this is *exactly* what I needed.
    
    Thanks a lot.
    
    Peter.
    

    For those people suffering from flat batteries, there is a command
    procedure
    
                              tbx:decnet.com
    
    which will do the area.nodenumber -> absolute address (and vice
    versa) for you. Try typing
    
                             @tbx:decnet help
    
    Steve

    Thanks Steve.
    
    [Thinks; wouldn't it help if I got some VAX training.....
    
    Must speak to my manager again.....]
    
    Peter		:^)
    

    I don't know whether anyones else has noticed this, but ........
    
    When logging in to the ADG cluster, the NOTICE tells us that RGDE28
    now has a full DECNET database .......... HOORAY !!!!!!!!!!!!
    
    However, when trying to reply to a mail message, logged in on RDGE28,
    I kept getting the "node not currently reachable message" .....
    ..... until I logged in to RDGE00.
    The node I was trying to access is in the USA and was involved
    in a recent change of area.
    It seems that RDGE00 has that area change reflected in the database,
    but RDGE28 does not. On investigation, I found that at least 1 other
    area and a few moved nodes are similarly affected.
    
    Moral of the story :
    
    Be warned .... things may not always be as they appear.  ;-)  ;-}
    
    Bob

    The DECNET route to the outside world for U39 is currently via U3
    and U28. This will change on Wednesday 7th Oct. U39 will be
    disconnected from U3 at 10:30 and reconnected to ADGV01, a VAX11/750
    designed to handle DECNET traffic only.
    
    The work will take approximately 1 hour. 
    
    U3,MAIDEN, and SEPTIC will be switched over to ADGV01 during the
    following few days. 

    Maiden and Septic will have their DMR11 cabled re-routed to ADGV01
    starting at 10:30 on Thurs 8th Oct. The work should take approx
    2 hours - DECNET to the PDP's will be unavailable while the work is
    in progress.
    
    U3 will have it's dmr cables rerouted starting at 3:30pm on Friday
    9th Oct. U3's DECNET will be unavailable for about an hour.
    
    

    I am *still* concerned about the lack of co-ordination between the
    network databases on the ADG cluster. It seems that RDGE00 reflects
    the current area/node numbers, but RDGE28 still has not been upgraded
    to reflect changes mode quite a while ago. Unfortunately, FOOT seems
    to be a copy of the database on RDGE28 and is therefore not right !
    Can this please be sorted out ? It gets very annoying sometimes,
    especially when trying to host to another node for support purposes.
    
    Bob
    

    yes