This problem literally spawns a decade so I made a story out of it to
create a better picture :-)

The story:
Back in 2002-2003 we had a server called 'server3' that was old, slow
and in its end was corrupting our NDS. It was an evil server.
When it started doing this we pulled the server and spent a week trying
to get our tree back in shape, doing all sorts of things with big signs
on it saying 'don't do this w/o tech support telling you to' (we didn't
have support at the time) like destroying whole replicas from all but 1
server and re-syncing it.
We got everything back in order (we thought) and about 6 months later we
added a new server with an identical name, result: all hell broke loose.
partitions refused to sync and what not. Pulled the new server, renamed
it, re-added it and everything was ok.
In the following 10 years we had very few issues with eDir, 1 or 2
crashed servers(hardware), some stuck obituaries and NW5 licenses that
could not be deleted after switching to all Suse but nothing major.
During these 10 years we always joked not to add a server called
'server3' while going from 3 servers to 10.
All our current servers are non ID migration servers, there is 1 that
still uses a name that was used back then but it did not have its old
identity migrated. The user containers from back then have since all
been replaced (except for 'servers' and an old zen application

The current problem:
A few weeks ago I added a new server to host the latest IDM
bundle(OES11sp1), that's when the trouble started.
People who authenticated to that server occasionally did not get their
drive mappings(I was thinking SLP issues for this one), but today I
managed to catch an error stating 'server3 not found'
So something somewhere somehow is still pointing at this old server, and
I can't find it ..

Does anyone have any suggestions on how to finally put this old server
to rest and purge it from our tree ?

Conz's Profile:
View this thread: