Dave,
Is the patch you mentioning here for filtering out callback transport?
Fong
-Original Message-
From: Dave Shield [mailto:[EMAIL PROTECTED]
Sent: Wednesday, August 25, 2004 2:28 AM
To: [EMAIL PROTECTED]
Cc: Fong Tsui
Subject: Re: Hosts.allow/hosts.deny relates to snmpd
Dave> I'v
It is broken for using multiple-word community name.
An example:
rwcommunity "A B" default in snmpd.conf
`snmpwalk -c "A B" localhost ucdavis` will get time-out because of no-matched
community name.
Attached is a fix patch against v5.0.9.
Thanks,
Fong
multi-word-community.patch
Descriptio
To use DISMAN-EVEMT, I configured with it and added the following fields in snmpd.conf
rouser me
monitor -u me -r 1 -o sysUpTime.0 -o memErrorName.0 "high process memory" memAvailReal
< 1000
Do I suppose to get trap? Apparently, no traps are received. Anything wrong here? I am
running on 5.0
On Wed, 2004-08-25 at 17:41, Robert Story wrote:
> On Wed, 25 Aug 2004 16:48:15 +0100 Dave wrote:
> DS> RS> I'd vote for the / style (more C99 headers).
> DS> RS> I'd like to avoid project specific types...
> DS>
> DS> OK - so what would that mean in practise?
> DS> What should the typedef name be
--with-mib-modules="disman/event-mib"
>I add disman in configure by --with-mib-modules="disman". But configure
complains:
error: module files disman.h or disman.c could not be found in
./agent/mibgroup
Do I need to copy agent/mibgroup/disman/event-mib.h to
agent/mibgroup/disman.h?
This communi
On Wed, 25 Aug 2004 16:57:13 +0100 Dave wrote:
DS> The libwrap calls check against "snmpd" if there's an address string,
DS> but "snmp" it there isn't. Is this difference intentional? If so, why?
Well, that change can be tracked to John:
http://cvs.sourceforge.net/viewcvs.py/net-snmp/net-snmp/
Dave,
I add disman in configure by --with-mib-modules="disman". But configure complains:
error: module files disman.h or disman.c could not be found in ./agent/mibgroup
Do I need to copy agent/mibgroup/disman/event-mib.h to agent/mibgroup/disman.h?
Thanks,
Fong
-Original Message-
From
On Wed, Aug 25, 2004 at 04:48:15PM +0100, Dave Shield wrote:
> DS> I'd suggest that the first thing would be to make sure that we have
> DS> suitable 32-bit types available - either defining new project-specific
> DS> types, or pick on a "standard" one and make sure it's defined
>
> RS> I'd vote f
On Wed, 25 Aug 2004 16:48:15 +0100 Dave wrote:
DS> RS> I'd vote for the / style (more C99 headers).
DS> RS> I'd like to avoid project specific types...
DS>
DS> OK - so what would that mean in practise?
DS> What should the typedef name be for:
DS>
DS> a) unsigned 32-bit integers
DS> b)
On Tue, 24 Aug 2004 23:32:22 +0100 Richard wrote:
RM> > Checking a few of the sourceforge compile farms hosts reveal that neither
RM> > FreeBSD or OpenBSD have SCNuMAX defined in . So instead of an
RM> > out-and-out replacement, we need to break down and use an ifdef on
RM> > __STDC_FORMAT_MACROS t
DS> I've just applied a patch to the main development code which should
DS> help with this situation (appended). Try that and see if it helps.
RS> If it works, it should probably be applied to the 5.1 and 5.0
RS> patches branches as well.
I'm getting forgetful in my old age - it appears that I
DS> I'd suggest that the first thing would be to make sure that we have
DS> suitable 32-bit types available - either defining new project-specific
DS> types, or pick on a "standard" one and make sure it's defined
RS> I'd vote for the / style (more C99 headers).
RS> I'd like to avoid project specif
On Wed, 25 Aug 2004 09:06:01 +0100 Dave wrote:
DS> Richard> All of interface.c uses ulong and long for the values. Hence if
DS> Richard> the kernel reports > (2^32)-1 then you get that value reported
DS> Richard> upwards.
DS>
DS> Ummm...
DS> What would happen if the "snmp_set_var_value" API call (
> NETSNMP_DS_LIB_APPTYPES seems to be the sum of all types of config that is to
> be read while
> NETSNMP_DS_LIB_APPTYPE seems to be the 'natural' type that is parsed.
Yes - that's right.
> but I do agree that I do not understand the exact implications of them.
No - neither do I :-)
That tend
-Coders,
how is a non-root "agentuser" intended to work together with persistent
storage (usually /var/net-snmp/snmpd.conf)?
AFAICS, 5.1.2 snmpd creates /var/net-snmp/snmpd.conf as *root*
*before* switching to the "agentuser" uid. When shutting down, it
consequently fails to save the persistent dat
On Wed, Aug 25, 2004 at 09:38:52AM -0400, Robert Story wrote:
> On Wed, 25 Aug 2004 09:55:43 +0100 Dave wrote:
> DS> I'd submitted the changes for the first of these, but didn't think
> DS> that I'd touched the other two. So I was a little suprised when
> DS> Magnus reported this problem - partic
-Coders,
has anyone recently tried embedded perl in the agent with 5.1.2?
Even with a simple
perl do "/path/to/perl_module.pl";
(i.e. with the example "hello world" script) in snmpd.conf, I get
a *weird* line in the log ("snmpd -Dperl"):
perl: initializing perl (/path/to/net-snmp-5.1.2/shar
On Wed, 25 Aug 2004 09:55:43 +0100 Dave wrote:
DS> I'd submitted the changes for the first of these, but didn't think
DS> that I'd touched the other two. So I was a little suprised when
DS> Magnus reported this problem - particularly when I searched my
DS> code and didn't find any mention of "age
Hi,
I have found an error in disk.c . On line 661 is a calculation of a
multiplier which is of type float.
But the calculation happens with integers. Therefore if the result is lower
than 1 you will always get 0 (zero).
The effect is that you get for the absolut sizes on disks zero instead of
the c
On Wed, 25 Aug 2004, Dave Shield wrote:
Dave> I've just applied a patch to the main development code which should
Dave> help with this situation (appended). Try that and see if it helps.
Robert> If it works, it should probably be applied to the 5.1 and 5.0
Robert> patches branches as well.
That's
[ First - *please* don't mail me privately, without copying
any responses to the mailing list. I don't have the time
or inclination to offer private, unpaid, SNMP consultancy.
Keep discussions to the list, where others can both learn
and offer advice. T
> this afternoon i do sometest , i found that
> first, If u want to sent v1 trap,u just need to set trap=6
> and type=any num you want,
Which API call are you using?
send_easy_trap?
send_trap_vars?
something else?
Note that you still need a "trapsink" or "trap2sink" entry
in your s
Dave> I've just applied a patch to the main development code which should
Dave> help with this situation (appended). Try that and see if it helps.
Robert> If it works, it should probably be applied to the 5.1 and 5.0
Robert> patches branches as well.
That's what I'd intended, but I wanted(/hope
> In example.c file, when you change 1.3.6.1.4.1.2021.254.8.0 into 1,
> the agent should send v2 trap,but in fact there is nothing send out
Hmmm...
What happens if you run the agent using
snmpd -f -Le -Dexample
and then set 1.3.6.1.4.1.2021.254.8.0 to 1.
What do you see?
Dave
---
> I think I'm going to wait on Dave for this one, as I believe I may have
> been barking up the wrong tree with my patch, and led you astray.
>
> My patch made the config code parse the same file with different types.
> I think the idea is actually to parse different files (eg snmpd.conf &
> agen
> I know you can send v1 and v2 trap in the agent
> but how to send v3 trap in angent?
>From the FAQ:
Where are these traps sent to?
-
SNMPv3 destinations can be configured using t
Does anyone have any thoughts on this?
Every message I send to the net-snmp-* list seems to be ignored..
* Dan Sully <[EMAIL PROTECTED]> shaped the electrons to say...
I'm attempting to query a Cisco switch, using the CISCO-STACK-MIB.
If I use snmpwalk on the command line, I get back something li
Bruce> Of the two, I find the ":" more intuitive.
Dave> Yes - that's the style that Robert and I had come to an agreement on.
Robert> I remember that one of my reasons was that '/' was a path separator.
Robert> As I was reviewing this thread, I realized that ':' is the path
Robert> separator for
> Which patch has this DISMAN-PING-MIB implementation?
The one entitled
SOURCECODE OF DISMAN IMPLEMENTATION
(patch #1014306)
Note that this will probably be included with the 5.2 code release,
but will *NOT* be configured in by defau
> Sorry, the leak is not in sedning trap. It is in event log in RMON.
> Attached is the fix. It is for 5.x version.
OK - Thanks for that.
It has been applied to all three branches of the code.
Please note that the Rmon code in this package isn't really supported,
not least because it works with r
I know you can send v1 and v2 trap in the agent
but how to send v3 trap in angent?
---
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail
Richard> All of interface.c uses ulong and long for the values. Hence if the
Richard> kernel reports > (2^32)-1 then you get that value reported upwards.
Ummm...
What would happen if the "snmp_set_var_value" API call (and similar)
applied a 32-bit mask to any integer-based value that it received?
>
> So, the value of { 0 0 } is appropriate, and NOT A BUG.
> The NETSNMP code MUST NOT be changed to return another value!
Don't worry, David. I have no intention of doing so.
It's one of the few areas of this agent that I'm 100% confident about :-)
Dave
-
33 matches
Mail list logo
