Systems :
- NW65SP7 & Post Patches
- NCS & Native
- Virtualised in VI3.5 ( Native Systems ) and physical NCS
- attached to differnt SAN-boxes over different FC-switches
- LAN switches are DELL 6224f/6224

Problem :
A copy job of a large image file (about 2G) starts with 30-40kb/s and
drops down to a few hundred kb/s. If we wait an hour and repeat the job
the performance is stable. We wait again and the problem reoccurs.
GW-Users complain that jumping from one GW-Folder to another can take up
to 50 sec. The auto completion for eMail addresses takes a similar time.
Migrating the GW-resource to the other node solves the problem for some
hours. This seems to be the same problem like the copy job has.
Health monitor in NRM shows all stats ok. Level II oplock are set to on
and client file caching is set to off.
One NW server is used as test system without antivir and backup agent (
pure NW-OS and Patches ).
Because we have physical and virtual servers the nic drivers seem not to
be the problem. Windows systems do not show any performace issues.
Until now we found no way to isolate the problem.