|
After spending the morning at Frontier, here's what I have to help.
First, there should be a file called l-r-user in the nshome directory.
This stands for logging and reports. It tells how to do both from
the command line. Logging must be set up using the command line,
but reports can be kicked off from within PROBEwatch once logging
is setup and running. You configure logging by editing nslog.cfg
Valid intervals for reports are 15 and 30 min, 1, 2, 4, 8 and 24
hours. The counters in the probe are 32 bits in length. It is
quite possible that some of them (e.g. octets) will wrap on a busy
network. You should sample often enough to avoid having them wrap.
Reports will detect one wrap, but can't tell if they wrapped twice.
The conversation report cannot detect if they have wrapped once.
pwmanager will start up the logging daemon, nslogd, when it starts up.
nslog.ctl will have the process id of the currently running daemon.
nslog.log is the log file for the logging; check it to see if logging
is running.
Log files are placed in the nshome/db/probename/domainname directory.
The file names are coded by year, month, day, hour.minute followed
by a letter E, T H, or C for Ethernet stats, Token Ring stats, Hosts,
or Conversations. The data is binary, but not compressed.
You need at least 2 files to run reports, since reports calculate
the delta between values in subsequent uploads.
If you get an error "nsrpt: No reports available for agent xxx",
it probably means that no data was logged or available for the
report to run. Reports are placed in nshome/reports.
I have placed 3 files in the NAC::NIPG:[MANAGE.RMON] directory:
logreprt.txt crontab.txt and dvadmin.txt
logreprt is a copy of l-r-user, which comes with the kit
crontab contains examples of how to run reports automatically
at defined intervals.
dvadmin.txt describes how to use the dvadmin utility to add more
TRAP destinations
|