I have 5 NW6.5 servers at the SP7 patch level. I have installed Security
Services 2.05 and eDir SP9 FTF3 (after the tree root was upgraded.)
Four of the servers have replicas. The fifth server is an iSCSI target and
has no replica. Of the four, two are cluster nodes, one is a ZENworks server
(backup also runs against this server) and one is for future use. The
primary cluster node exhibits the worst memory growth in these modules, then
the second node, ZENworks and future use. Even with no services running on
it, the future use server withers away. The storage node never has any
change in these modules. Prior to applying SS2.05 and the eDir patch, the
servers exhibited periodic growth in the problem modules. Since applying
them, the servers exhibit constant growth. I've included segstats and config
for the primary node and the ZENworks server. As you will see in the
segstats for the ZENworks server, in the 3 hours between when I reset the
server and started seg, the problem NLMs had already grown significantly at
6 and 9MB, respectively. After 1 day, they are up to 42/69mb. On the primary
cluster node, XENGEXP is over 180mb after a day and a half. The cluster
nodes and ZENworks server are pretty useless after about 3 days and need to
be reset. Any guidance will be much appreciated.

Erick Larson
Perkins Library