I would like to start a discussion off regarding what people feel are
the optimal settings for Netware 6.5. I appreciate that everyones
circumstances are different but the set parameters are a minefield and
maybe a good starter for 10 might help many starting out at this
version.

Through using Netware 6.5 for 18 months+ now and through trial and
error and also looking at other peoples experiences on such forums as
this I have come up with the folloiwng SET changes:-

SET CLIENT FILE CACHING ENABLED = OFF
set cpu hog timeout amount = 7 sec
set directory cache allocation wait time = 1 sec
set directory cache buffer nonreferenced delay = 30
SET DIRTY DISK CACHE DELAY TIME = 0.5 SEC
set enable hardware write back = on
SET FILE SERVICE MEMORY OPTIMIZATION = 0
set garbage collection interval = 5 minutes
SET LEVEL 2 OPLOCKS ENABLED = OFF
set maximum concurrent directory cache writes = 500
SET MAXIMUM CONCURRENT DISK CACHE WRITES = 2000
set maximum concurrent opens per file and connection = 1000000
set maximum directory cache buffers = 200000
set maximum number of directory handles = 1000
set maximum number of internal directory handles = 1000
set maximum outstanding ncp searches = 500
SET MAXIMUM PENDING TCP CONNECTION REQUESTS = 4096
set maximum record locks per connection = 10000
set maximum wait states = 20
set minimum directory cache buffers = 1200
set minimum file cache buffers = 200
set ncp protocol preferences = tcp udp
SET NLS SEARCH TYPE = 1
set read ahead lru sitting time threshold = 60
set sequential elevator depth = 6000
SET TCP DELAYED ACKNOWLEDGEMENT = OFF
set tcp nagle algorithm = off
set tcp ip maximum small ecbs = 65534
set tcp minhsall algorithm = on
set vm garbage collector period = 5 minutes

These are the only ones we change. Some are a throw back to the 5.1
days. Anybody like to comment or add their own suggestions.

I can highlight what a dramatic effect some of these parameters can
have. The 'set garbage collection interval' was always set to 1 minute
prior to SP5. When left at 1 minute this parameter caused a high
utilisation of a processor to occur. Simply changing it to 5 minutes
stopped this. We also noticed that it fixed an ongoing problem with a
server that has SP3 on it and which had response tiem issues. For
example when set to 1 minute:-

Ping Statistics Tuesday 17th January 2006 about 10.30
==================================
Ping statistics for xxx.xxx.xxx.xxx
Packets: Sent = 1000, Received = 913, Lost = 87 (8% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 2986ms, Average = 44ms

When set to 5 minutes:-

Ping Statistics Tuesday 24th January 2006 about 15.30
=================================
Ping statistics for xxx.xxx.xxx.xxx
Packets: Sent = 1000, Received = 1000, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 85ms, Average = 1ms

I would be interested to hear other peoples experiences with other
parameters which I could collate.

Chris