Help on SNMP V3 and host downed

[cere23]

Hi to all.

I have a problem during upgrade from 0.8.7a to 0.8.7b.
I'm using cacti for monitoring a large infrastructure ( about 1000 hosts ) using cacti plugin (thold and monit). Some hosts are confgured in SNMP v 2c and some other in Snmp v3.

After upgrade all seems to work fine but today I realize that cacti start not detecting hosts down . If i use snmp v 2c cacti work fine.
I use snmp to probe hosts downed.

Any help will be appreciated.
Best regard and tanks in advance
(Sorry for my bad english)
Andrea

[gandalf]

To sum up:
Using SNMP V3 shows hosts as up even if they are down, correct? WHich SNMP V3 parameters do you use?
Reinhard

[cere23]

Yes sum is correct.
Today I'm not in office but the configuration is something like that:

syscontact admin@root
syslocation Some_place
rouser admin auth
sysservices 15


All hosts are Solaris 5.8 with net-snmp 5.4.1 compiled from source.

When the host is up snmpwalk works fine.

tanks in advance

[gandalf]

Oops, Solaris. Somewhat special. Do you use cmd.php or spine? I suppose, spine. When running spine against the "failing host", do the results differ from running cmd.php against this host?
Reinhard

[cere23]

Currently l'm using cmd.php because of spine give me back too much false positive. So I can't use spine although cmd is less performant .

Any other idea ?

tanks
Andrea

[gandalf]

ATM, I'm only interested in running both cmd.php and spine against one of the failing hosts from cli. You may stick to the current poller but run those commands from cli only. Both allow for specifying a specific host only.
I'm interested, whether output differs
Reinhard

[cere23]

If I understood i must run from the cli of my cactibox

php cmd.php hostname

and

spine hostname

and report you the result ?

TNX

[gandalf]

it's
php -q cmd.php <first host> <last host>
and
spine --verbosity=5 <first host> <last host>
Output of the first one is found in cacti.log only whereas the latter one prints to stdout as well. cmd.php uses the verbosity given from "Settings"
Reinhard

[cere23]

Hi, You have got the output in attached file.

Tank you

Andrea

[gandalf]

You did not enter the same hosts for cmd.php and spine, did you?

From cmd.php

04/13/2008 08:23:23 PM - CMDPHP: Poller[0] Host[727] SNMP: Device does not require SNMP
04/13/2008 08:23:23 PM - CMDPHP: Poller[0] DEBUG: SQL Exec: "update host set status = '3', status_event_count = '0', status_fail_date = '0000-00-00 00:00:00', status_rec_date = '0000-00-00 00:00:00', status_last_error = '', min_time = '0.00000', max_time = '0.00000', cur_time = '0', avg_time = '0', total_polls = '15934', failed_polls = '0', availability = '100' where hostname = 'xxx.xxx.xxx.xxx'"
04/13/2008 08:23:23 PM - CMDPHP: Poller[0] DEBUG: SQL Assoc: "select poller_reindex.data_query_id, poller_reindex.action, poller_reindex.op, poller_reindex.assert_value, poller_reindex.arg1 from poller_reindex where poller_reindex.host_id=727"
04/13/2008 08:23:23 PM - CMDPHP: Poller[0] Host[727] DS[14696] SNMP: v3: xxx.xxx.xxx.xxx, dsname: ucd_memTotalReal, oid: .1.3.6.1.4.1.2021.4.5.0, output: 2097152

shows Host[727] with a strange behaviour. But spine was called for Host[10]

04/13/2008 07:45:11 PM - SPINE: Poller[0] DEBUG: SQL:'SELECT action FROM poller_item WHERE action=2 AND host_id BETWEEN 10 AND 10 LIMIT 1'
04/13/2008 07:45:11 PM - SPINE: Poller[0] DEBUG: StartHost='10', EndHost='10', TotalPHPScripts='0'

Please clarify
Reinhard

[cere23]

Yes, It's the same host polled in different time.
I used cli for spine and gui in for cmd.php in debug mode.

Bye
AC

[gandalf]

Yes, It's the same host polled in different time.
I used cli for spine and gui in for cmd.php in debug mode.

Bye
AC

??? How can Host[727] = Host[10] ???
Reinhard

[cere23]

I'm sorry ..
I've attached the correct file
Tanks
bye
Andrea

[cere23]

File