Netreo SNMP notes

2005-06-24:

http://omniscient1.nac.uci.edu/ata/summary_view.php

Scroll down to "DCS Servers" and click it

Scroll down to the host you want to edit, and click it

Near the upper middle of the screen, there should be an "edit" link - click it

Click Polling

Wipe out the "write community", and fill in the "read community"

Click "apply changes" at the bottom of the screen

Click "quick SNMP test" - when I click it, I get no SNMP traffic, but
I do get some HTTP traffic :-S

This works though:
seki-strombrg> snmpwalk -v 2c -c xxxxxxxx nsc-1.nacs system.sysDescr
SNMPv2-MIB::sysDescr.0 = STRING: Linux
network-stats-collector-1.nacs.uci.edu 2.4.21-32.0.1.EL #1 Tue May 17
17:55:54 EDT 2005 i686

It's beginning to seem like Netreo supports v1 and v2, while RHEL's
net-snmp supports v1, v2c and v3.  I had an RHEL3 v2c config file that
would work with FC4's snmpwalk, however, it does not appear to work with
omniscient's v2.  I then changed my RHEL 3 snmpd.conf to v2, but FC4's
snmpwalk could no longer walk the RHEL 3 box.  Same was true with v1 in
the RHEL3 snmpd.conf - there must be more to changing SNMP versions in
RHEL3 than just changing the "sec.model" entries.

Something phase of the moon may be going on - I just restart snmpd on
RHEL3 again, and smnpwalk from FC4 with version 1 is working.

snmpd.conf had omniscient1.nacs, but should have had omniscient1.nac.
Upon fixing that, omniscient1 could query v1 using net-snmp v1, and
could query v2 using net-snmp v2c.

I'm able to get system.sysDescr, loadaverage, and more in the simple
snmp test under polling.

I've now moved on to the advanced snmp testing.  I experienced another
seemingly phase of the moon issue: snmpwalk on seki could not get any
snmp data at all.  service restart snmpd again "resolved" the problem.

An snmpwalk from seki to nsc-1 via v2c gives the following entries
including the string "yp" but not including the string "type":
HOST-RESOURCES-MIB::hrSWRunName.1630 = STRING: "ypbind"
HOST-RESOURCES-MIB::hrSWRunPath.1630 = STRING: "ypbind"
HOST-RESOURCES-MIB::hrSWRunParameters.1726 = STRING: "--timeout=60 /Dcs
yp auto.Dcs -rw,intr,hard,nosuid,timeo=20,rsize=8192,wsize=8192"

-HOWEVER-, I cannot run that same command twice in a row on seki, unless
I restart snmpd on nsc-1 in between the two snmpwalk's.

Trying up2date on nsc-1, in case there is an already-released snmpd fix...
up2date is getting stuck on a futex thing, and as it happens, so is snmpd.
Rebooting...

This is from snmpd, prior to the reboot:
futex(0xb7265250, FUTEX_WAIT, 2, NULL <unfinished ...>

Since the reboot, snmpd is no longer having this problem.  This may point
at a reliability problem.  Fortunately, the host is running RHEL 3, not 4.

Some interesting OID's:
01 command[check-host-alive]=$USER1$/check_ping $HOSTADDRESS$ 100 100
5000.0 5000.0 -p 1
02 command[check_mrtg]=$USER1$/check_mrtg $ARG1$ 10 AVG $ARG2$ $ARG3$
$ARG4$ $ARG5$ $ARG6$
03 command[check_users]=$USER1$/check_users $ARG1$ $ARG2$
04 command[df-500MB]=$USER1$/check_snmp -H $HOSTADDRESS$ -o
.1.3.6.1.4.1.2021.9.1.7.1 -w 5000000: -c 4000000: -C public
05 command[check-automount]=$USER1$/check_snmp -H $HOSTADDRESS$ -o
.1.3.6.1.4.1.2021.2.1.5.1 -w 1: -c 1: -C public
06 command[check-X]=$USER1$/check_snmp -H $HOSTADDRESS$ -o
.1.3.6.1.4.1.2021.2.1.5.2 -w 1: -c 1: -C public
07 command[check-kdm]=$USER1$/check_snmp -H $HOSTADDRESS$ -o
.1.3.6.1.4.1.2021.2.1.5.3 -w 1: -c 1: -C public
08 command[check-ypbind]=$USER1$/check_snmp -H $HOSTADDRESS$ -o
.1.3.6.1.4.1.2021.2.1.5.4 -w 4: -c 2: -C public

This too looks relevant, but may also not really be:
http://www.alvestrand.no/objectid/1.3.6.1.1.1.2.html