While working w/a user in the IRC channel today, we found a workaround for this
bug, which basically involved increasing a buffer so that it can hold all ip
addresses. Since we can't predict the number of interfaces (they had around
250, default buffer hold 20), I don't like the idea of just making
Robert Story wrote:
It was already in 5.2.x, and I've added it to 5.1.x. 5.0.x is in rc for 5.0.10,
and I don't think fixing a compiler warning is worth doing another rc, so I
left it out there.
I agree. But if there'll be a 5.0.10.rc3 for other reasons, I think this
should be added before, to
On Thu, 09 Jun 2005 19:00:50 +0200 Thomas wrote:
TA> Robert Story wrote:
TA> > In cvs MAIN, someone added a define of __EXTENSIONS__ in
TA> > net-snmp/system/solaris.h. This does not appear to have made it into
TA> > 5.0.x.
TA>
TA> This should probably be added to 5.[012] as well.
It was already
-Original Message-
From: Robert Story [mailto:[EMAIL PROTECTED]
Sent: Wednesday, June 08, 2005 8:35 AM
On Wed, 08 Jun 2005 09:36:29 +0100 Dave wrote:
DS>> You might be able to short-circuit this by tweaking the
DS>> var_XXX() code, so that it updates the 'name' parameter
DS>> to refer to
[EMAIL PROTECTED] wrote:
NET-SNMP version: 5.2.1
[...]
Jun 9 12:02:01 scc-hs2001 snmpd[12465]:
/usr/local/share/snmp/snmpd.conf: line 79: Warning: Unknown token:
smuxpeer.
Sounds like smux support has been disabled in your agent.
- What does "net-snmp-config --configure-options" tell?
- W
hi thomas,
thx for the reply =)
after a successful build/install of either v5.2.1 or 5.3.dev on OSX
10.4.1, a check of operation with:
/usr/local/net-snmp/sbin/snmpd -v
dyld: Symbol not found: _boot_DynaLoader
This is related to Perl.
i _just_ got that ...
If you've enabled embedded p
I want to write a simple application to send traps (similar to the snmptrap app
from net-snmp). The idea is, that the application takes all trapsinks from the
snmpd.conf file and send the trap to all defined trapsinks.
Is there another (easier) way than to call get_configuration_directory() to
Hello
I have installed
NET-SNMP version: 5.2.1
Web:
http://www.net-snmp.org/
Email:
net-snmp-coders@lists.sourceforge.net
The lotus notes configuration requires
that I install lnsnmp and add an smux value to snmpd.conf
smuxpeer 1.3.6.1.4.1.334.72 NotesPasswd
Howeve
Hi Wes,
Thanks so very much for your response and it works well!
The Makefile generated is to produce a file that can be run as a daemon, the snmpd.
My question now is how to make modifications to the Makefile to produce an object library,
like an *.o file so that it will be linked to the pr
Mike Fisher wrote:
Bruce Shaw wrote:
Could you post it as a patch? I'd like to have a look at it. My code
works on my E450's, V880's and 280r's.
Minor clarification: I have code, but it's not been integrated back into
the agent yet. (hence my original query) It's also not working as well
OpenMacNews wrote:
after a successful build/install of either v5.2.1 or 5.3.dev on OSX
10.4.1, a check of operation with:
/usr/local/net-snmp/sbin/snmpd -v
dyld: Symbol not found: _boot_DynaLoader
This is related to Perl. If you've enabled embedded perl, watch the
corresponding build proc
Robert Story wrote:
On Thu, 09 Jun 2005 10:37:37 +0200 Thomas wrote:
TA> As you said, the question now is how to *properly* (best) fix our build
TA> system.
If the header definition isn't getting picked up, does this imply that net-snmp
is not detecting the native implementation, and is using
hi all,
after a successful build/install of either v5.2.1 or 5.3.dev on OSX 10.4.1, a
check of operation with:
/usr/local/net-snmp/sbin/snmpd -v
dyld: Symbol not found: _boot_DynaLoader
Referenced from: /usr/local/net-snmp/lib/libnetsnmpagent.5.dylib
Expected in: flat namespace
w
-Coders,
I managed to compile net-snmp 5.2.1 on Tru64 Unix 5.1B (Alpha EV5.6)
using Compaq C 6.5. FYI, here's what needed to be done:
- apply my build fixes (Patch #1217475)
- ./configure --with-out-mib-modules="mibII/tcpTable mibII/udpTable"
("host" doesn't compile either, see Bug #1217509)
On Thu, 09 Jun 2005 10:37:37 +0200 Thomas wrote:
TA> As you said, the question now is how to *properly* (best) fix our build
TA> system.
If the header definition isn't getting picked up, does this imply that net-snmp
is not detecting the native implementation, and is using the version in
snmplib?
We are using SNMPv2 and the ucd 4.2.5 version of the libraries
-Original Message-
From: Dave Shield [mailto:[EMAIL PROTECTED]
Sent: Thursday, June 09, 2005 1:05 AM
To: Toth, Gregory S
Cc: net-snmp-coders@lists.sourceforge.net
Subject: Re: Manipulating the reqid in a pdu
On Wed, 2005-06
Dave Shield wrote:
On closer examination, the agent seems to be falling over as soon
as it receives the initial SNMPv3 engine probe. It quits with
a SegV fault in 'usm_add_user_to_list', while trying to construct
the REPORT response.
Ring any bells with anyone?
Didn't we change code to delay
On Thu, 2005-06-09 at 13:47, Thomas Anders wrote:
> with a clean 5.3.dev CVS checkout, "make test" fails for me (on a fully
> patched SuSE Linux 9.3) on all AgentX, proxy and SNMPv3-related tests.
> Can others reproduce this?
Yup.
I'm getting the same thing.
On closer examination, the agent see
-Coders,
with a clean 5.3.dev CVS checkout, "make test" fails for me (on a fully
patched SuSE Linux 9.3) on all AgentX, proxy and SNMPv3-related tests.
On the same system, the 5.2 CVS branch succeeds with all tests.
Can others reproduce this?
+Thomas
--
Thomas Anders (thomas.anders at blue
Amber Gupta wrote:
Actually I looked at your patch earlier and had same opinion that you
are trying yo provide solution for same.
The problem with me is that I am relatively new to net-snmp and still do
not know how patch works..Do I need to just copy file an compile?
Get the patch (say: file.d
Andrew Hood wrote:
Thomas Anders wrote:
I'm seeing the same on Solaris 9 (w/ Sun cc). Looking into
/usr/include/string.h, the strtok_r declaration is nested inside:
#if defined(__EXTENSIONS__) || defined(_REENTRANT) || \
(_POSIX_C_SOURCE - 0 >= 199506L)
extern char *strtok_r(char
On Wed, 2005-06-08 at 18:42, Toth, Gregory S wrote:
> We are running an application where we manage retrys of snmp requests
> rather than defining the retrys in the session.
> we decided maybe if we could send all
> the retrys out with the same request id as the or
On Thu, 2005-06-09 at 02:38, Alex Burger wrote:
> Wes Hardaker wrote:
> > We should consider making a development exe too for people that want a
> > full installation so they can use the libraries... There is probably
> > enough users that want to do that to warrant creating it?
>
> Do we want a
23 matches
Mail list logo
