Set a lower threshold for the CPU temp.
2009/6/4 PoWah Wong :
>
> I am designing a monitoring program for the fan and CPU temperature in my FC6
> linux computer using lm_sensors and net-snmp 5.4.2.1.
> If there are fan and CPU temperature problems, then a snmp trap should be
> sent.
> How to tes
It looks like someone inserted MBG-SNMP-MIB's text in the middle of
http://www.net-snmp.org/wiki/index.php/TUT:Configuring_snmptrapd -
since it's a huge mib, this is a somewhat intrusive change. I removed
it. If for some weird reason this was actually supposed to be there,
feel free to reinsert it.
What are the exact strings you added to the config files? And did you
change the default config file path (eg. with --prefix) when you ran
configure before compiling?
-Dan
2008/12/8 BOMB <[EMAIL PROTECTED]>:
> I've ported the net-snmp v3 to a mips embedded system, after copied the
> libs
>
I asked this on the net-snmp coders list, but perhaps it's better
suited for here. There is a session->usmUser converter, but I can't
find a usmUser->session converter. It's not quite as simple as copying
over fields, since in a multilingual situation the snmpCommunityTable
might need to be referen
e a while.
>
> I guess that I have to set up snmpd.conf ??? If so, how to do it ???
>
> Thanks so much for your help
>
> Duy
>
>
>
> - Original Message - From: "Mike Ayers" <[EMAIL PROTECTED]>
> To: "dan anderson" <[EMA
snmptrapd is a trap listening daemon, and will not send traps.
notification.c & .h will compile into a module which sends a
notification/trap (netSnmpExampleNotification is the one) twice a
minute. So with the snmptrapd running, if you then compile the
notification example into the agent (snmpd),
ge-
>> From: [EMAIL PROTECTED]
>> [mailto:[EMAIL PROTECTED] On
>> Behalf Of dan anderson
>> Sent: Wednesday, September 10, 2008 3:53 PM
>> To: net-snmp-users@lists.sourceforge.net
>> Subject: Re: v2 notifications as v1 traps
>>
>> Nobody knows? Hmm. Would any
net-snmp snmpd or I'd do it myself, obviously.)
Cheers,
-Dan
2008/9/10 dan anderson <[EMAIL PROTECTED]>:
> When sending v2 notifications as v1 traps, is it correct to set the
> enterprise OID to the parent of the trap, the generic type to
> enterprise, and the spe
When sending v2 notifications as v1 traps, is it correct to set the
enterprise OID to the parent of the trap, the generic type to
enterprise, and the specific type to the last digit of the
notification's oid?
Eg. for ucdShutdown, would it be enterpriseOID: ucdavis.ucdTraps,
generic: 6 (enterprise-
Hi all,
objid_enterprise is hardcoded as .1.3.6.1.4.1.3.1.1 in the source
files. This seems to be a subtree of Carnegie Mellon's
enterprise-specific MIB. Why is this hardcoded to this? And why the
trailing .1.1? This is sent as the enterprise-specific OID when traps
are sent, so it's not a purely
Two questions:
Is it possible to run mib2c without perl modules installed? I get an
error and haven't found a workaround.
Either way, is it possible to use the resulting .c and .h files
without the perl modules? I expect so, since they're obviously C or
C++ files, but I figure it's best to check
In net-snmp/agent/mibgroup/target/target.c line 176 the comment
suggests that there's a memory leak. I can't figure out why - there's
no mallocs or news, param is the only other pointer returned by a
function, and it's owned elsewhere, and the netsnmp_transport_free
function doesn't have any note s
I've been hunting through the trap-related code, and it looks like
SNMP-NOTIFICATION-MIB and SNMP-TARGET-MIB aren't supported and the
related RFCs aren't followed. Is that correct? (That would mean that
sinks can only be registered using config files.)
I realize they can't be force-included and sh
Is there a tutorial for using table_iterator and other handler
helpers? http://www.net-snmp.org/dev/agent/examples.html has a few
very basic examples (eg. the 'watched' example registers a simple
handler), but even
http://www.net-snmp.org/docs/man/netsnmp_handler.html, though it
specifically mentio
Sure, you can do it like this (snagged from the host resources mib):
hrDeviceStatus OBJECT-TYPE
SYNTAX INTEGER {
unknown(1),
running(2),
warning(3),
testing(4),
down(5)
}
I believ
>Or just pull the freakin' cable!
Hard to pull a cable in vmware. ;-)
>Alternatively, use IP aliasing and mess with a network that doesn't
> have anything connected to it, like (not tested):
>
> $ ifconfig eth0:0 1172.16.1.27 netmask 255.255.255.0 # linkUp
> $ ifconfig eth0:0 d
On a somewhat related note, is there a way to force link up/downs to
occur on Linux, to test these traps?
-Dan
2008/6/17 Dave Shield <[EMAIL PROTECTED]>:
> 2008/6/17 k_simplydev <[EMAIL PROTECTED]>:
>>> Try shutting the agent (on the 20 system) down completely.
>>> Then run
>>>
>>> snmpd -f -
> It's not clear what you mean by "a standalone".
>
> The 'send_v2trap' routine is part of the agent API.
> If you're working with a standalone application, then
> this won't work. You'll need to use 'snmptrap.c' as
> a template in this case.
Ah, thanks. I didn't realize that API calls would req
In an attempt to send traps from a standalone, I've done the following:
create_trap_session("localhost", 0, "public", 1, SNMP_MSG_TRAP2);
init_traps();
[ populate notification_vars ]
send_v2trap(notification_vars);
But without success. When I step through in gdb, the sinks list is
empty. I've tri
Well, if you tweak the code and then commit it, the black box would
remain intact. ;-) Best to ask on the coders group if that's something
they'd appreciate, I suspect.
-Dan
2008/6/4 Ron Rader <[EMAIL PROTECTED]>:
> Dave Shield:
>
>> It shouldn't be too difficult to add this functionality,
>> so
No, by default SNMP is over UDP, which is connectionless. Trying to
send a UDP packet might throw or return an error if there's no network
interface to send it out on, or might do so pre-sending it if you try
to resolve a nonexistant name, but if a UDP packet gets out onto the
network it's gone and
If you unplug the cable, the message will not be relayed to NMS unless
you're using wireless. Or, I suppose, the Pigeon Protocol.
-Dan
2008/5/30 Murilo Fujita <[EMAIL PROTECTED]>:
> Hi!
> Sorry my last e-mail. Let me try again:
> I want to know if the trap was sent by the agent to th
Hey all,
I'm trying to compile net-snmp on x86 - one caveat being that I'm
compiling from a different machine than it'll run on. After some minor
haggling with the configure script and a whole lot of time reading old
posts to this list and random webpages (and applying the results) it
now compiles
23 matches
Mail list logo
