Hi all!

Having a DS replica problem:

History:

We wanted to replace an old BM server for a new NW6 + BM3.7 server.
After checking the two existing servers (THOMCAST-BM1 and EKT-SERVER),
I started to install NW6. Here I did an error: I gave the new server
(FIREWALL1) the same DMZ-address as DMZ on the old, still running
server, 192.168.5.1 ! Consequently, the installation hung in the
eDirectory install screen. DSTRACE=+s showed
"UpdateLocalTransitiveVector failed, replica not on (-673)"

Shut down the new server and wanted to do a reinstall.

When I checked the remaining servers with dsrepair, I noticed that ALL
THREE servers were „UP".

After some braining I changed the DMZ IP on THOMCAST-BM1 to
192.168.5.2, did a repair Network Address, but all servers are still
„UP"

I have not yet tried to delete FIREWALL1 in NDSMGR.
I think two servers in ds are sort of still pointing to the same
physical server (THOMCAST-BM1) and if I now try to delete FIREWALL1 in
the DS, NDSMGR might destroy both replicas, creating a havoc!

in replica ring all servers are present

timesync ok: „Time is in sync" = Yes on ALL servers (including
FIREWALL1, which isn't running at all)

„Servers known to this database": ALL servers have „Local Status" = UP

Network Address for Firewall1 (Server Information in NDSMGR) points to
the DMZ-Adapter on the THOMCAST-BM (192.168.5.1)!!
Did a „repair all network addresses" – error 634 no change, FIREWALL1
is still „UP" !!

Users can still login, but I really need to clear up NDS.


EKT-SERVER: [Root] Master – ON, NW5.1 SPACK 5.1.5 (V5.0 SP for Netware
5.1), DS 8.82

THOMCAST-BM1: [Root] Read/Write – ON, NW5.0 SPACK 5.0.6 (V6.0a SP for
Netware 5), DS 7.51

FIREWALL1: [Root] - New Replica - ON, Server not running (DSREPAIR
show DS V 7.51)

In NWADMIN 2 new Objects – Server FIREWALL1 und Unknown Object
FIREWALL1-PS


Should I delete the server object? Repair local ds?

lars

-