since NW6.5SP3 (->OES) one of our clusters seems to eat up memory. It runs
fine for a day or two and then it allocates memory to the miscellaneous
memory within hours, leaving nothing to live. volumes become inaccessible
This happens on both nodes, most of the time on one special nodes.
I have checked and tried all the tids regarding memory fragmentation and out
of memory messages, restricted NSS-memory with file cache maximum size,
restricted DS to a hard limit, but still we get these situations.
The cluster has ZfD, GW, DNS, SLP, legato.
The other cluster that is doing Bordermanager it doesn t happen.
I believe it must be something with the Modules, but everything is
up-to-date. There is even nothing in Beta that sounds interesting.

help would be appreciated

kind regards

Gerd Zobel
The NetWorker GmbH