I have an older IBM Netfinity server (4000R) with 2 Pentium III chips and
a GB of RAM. It ran flawlessly for years with Netware 6.

Recently, we installed OES (with the SLES 9 linux kernel) for a web based
FTP server. For the most part, everything runs fine....for about 2 days.

At first, it just crashed. Black screen and non-responsive. I passed
NMI_Watchdog=1 to the kernel at boot up and the next time it crashed I
saw in /var/log/messages that the server killed the novell-xsrvd daemon
about 10 to 15 times (said the kernel was out of memory and from what
I've read NMI_watchdog only kills processes that lock the CPU at 100% for
more than 5 seconds).

Anyways, I found a TID that states you need to add a path to the
LD_LIBRARY variable. I've applied this fix and I don't have problems
runnng xrvcfg. However, I still notice that a few times a day the
processors run at about 97% for 30 minutes or so. When this happens there
are 10 novell-xsrvd processes running (each at about 19% CPU utilization).

Is this normal? The rest of the day it runs at about 0.5 to 2%