On a number of my NW6.5SP7 (and earlier) servers, memory allocated to
xengexp.nlm and xmgr.nlm periodically begin to grow.This is particularly bad
on my cluster nodes. These modules grow in size until there is no available
memory and the server requires a reset. It appears that this issue is
related to the NDS master server and/or the root SLP servers. If some (or
all) are reset at the root, the issue does not arise for a while. If nothing
is done at the root, resetting my servers fixes the immediate problem but
they start growing again. Of 5 servers, 4 exhibit the problem (with the
cluster nodes the worst) and one never does. The one that does not have the
problem does not contain a replica. The 4 problem machines do. Others in the
tree, particularly those with clusters, also have the same issue. We cannot
find anything on the issue in support docs. Any clues?

Erick Larson
Perkins Library
Duke University