I'm running OES Linux SP2, fully patched via rug, eDirectory I have
a 7.5TB NSS volume on which our users each have a 10GB user quota. The
server is a 2x3GHz Xeon, 2GB RAM, dual 1000Mb ethernet.

Normally the drive maps almost instantly, but several times throughout the
day, we see drive mappings to that volume become very slow. Meanwhile
mappings to our others servers are as fast as they always are. Additionally,
either NCP or NSS hangs every few days. Connections to the volume on that
server become disconnected, the drive mappings to that server fail, and we
end up restarting eDirectory on that service to restore service.

We have about 1000 or more users connecting, but all of the users also have
mapped drives to a Netware server and we never see any issues with that
server. When the slowdowns or hangs occur, there's no indication that the
server is starved for resources. Neither CPU, RAM, nor disk or network IO
show heavy usage. No CPU hogs. NCPCON stats may show 900 NCP connections and
perhaps 5-10 open files.

Other info: shredding and compression are off, immediate purge is on.

I've increased n4u.server.max-threads from 64 to 256 threads.

I'd be grateful for any help you can provide in either diagnosing or
troubleshooting this. If you need more info, just let me know. Thanks!