We're in the process of shifting a few things around and I want to
remove a read/write replica from one of our servers.

From past experience I have found that doing this causes users logins to
become stuck. Users will log out of their workstations and it looks
like all connections were closed but the network address value for the
machine they were using at the time the replica was removed remains
present. This results in the "You are trying to log on to too many
workstations" message as we generally have user's concurrent logins set
to 1.

The only ways to clear it appear to either do it manually (using LDIF)
or to increase their concurrent logins to 2 and get the user to log in
to the same workstation they were using. If they use a different
workstation the old network address value remains present but will be
removed if they use the same workstation.

I've managed to re-create this behaviour in a test environment so it
doesn't appear to be a one-off and I was wondering what I can do to
avoid this problem.

OS: SLES 10 Sp4, OES2 Sp3
eDir: 8.8 Sp6

(Old I know, we're planning upgrades/replacements hence this post).

Any advice would be appreciated.



reddragon27284's Profile: https://forums.netiq.com/member.php?userid=6842
View this thread: https://forums.netiq.com/showthread.php?t=52392