🚀 go-pugleaf

In article <1993Jan8.215322.16191@riacs.edu>, evans@pan.arc.nasa.gov (Dan Evans) writes:
 
| I hope this will be an easy question for some of you to answer.
| I'm no DECnet wiz, so the following is my perceptions of what is going
| on, please excuse me if I'm wrong.
| 
| We have a LAVC running VMS 5.5-2.  It is a single system disk cluster.
| The NCP database (for remote hosts) is stored in a common area
| 	SYS$COMMON:[SYSEXE]NETNODE_REMOTE.DAT
| so all the systems can share it.  (Or so the VMS Networking (Vol 5A)
| manual claims on pg 5-42.)
| 
| The problem I'm having is, whenever I update the database on one node in
| the cluster, the new information is not reflected on the other nodes.
| Even though they all use the same file to load their database.
| 
| For example,  here I change a node name to reflect a change in address:

[ example omitted ]

Dan,

You should make your change(s) to the permanent database via an
NCP PURGE/DEFINE on a single node.  Then use SYSMAN to update everyone's
volatile database using NCP CLEAR/SET.

So if you want to get rid of node XYZ, then re-add it with addr 5.10 (per your
example):

$ MC NCP
NCP> PURGE NODE XYZ ALL
NCP> DEF NODE XYZ ADDR 5.10
NCP> ^Z
$ MC SYSMAN
SYSMAN> SET ENV/CLUS
SYSMAN> DO MC NCP CLEAR NODE XYZ ALL
SYSMAN> DO MC NCP SET NODE XYZ ADDR 5.10
SYSMAN> ^Z

I hope the above has been helpful.  If you need further assistance,
please feel free to call Bob Gezelter. :-)

Aaron

Aaron Leonard (AL104), <Leonard@Arizona.EDU>
University of Arizona Network Operations, Tucson AZ 85721

In article <1993Jan8.215322.16191@riacs.edu>, evans@pan.arc.nasa.gov (Dan Evans)
writes:
>Why is this happening?  

There are two databases permanent (in disk) and volatile (in memory). When
DECnet starts up, the permanent database is used to initialize the volatile
database. When you used SET NODE you updated the volatile database on the node
you are on. When you used DEFINE NODE you updated the permanent database. But
you didn't update the volatile database on any other node. You need to either
restart DECnet on each other machine (yech), or issue the following commands on
each other node:

NCP> clear node xyz
NCP> set node xyz all

The first command clears the node from the volatile database, the second copies
the node info from the permanent database into the volatile database. So to do
it on every node, try something like this (untested):

$ NCP
NCP> purge node xyz all
NCP> def node xyz  512
NCP> exit

$ mc sysman
SYSMAN> set environment/cluster
SYSMAN> do ncp clear node xyz all
SYSMAN> do ncp set node xyz all
SYSMAN> exit

Apologies if any of this is wrong. I'm sure someone will say so. :-)
--
Terry Poot <tp@mccall.com>                   The McCall Pattern Company
(uucp: ...!rutgers!depot!mccall!tp)          615 McCall Road
(800)255-2762, in KS (913)776-4041           Manhattan, KS 66502, USA