Error messages in snmpd.log

2010-08-23 Thread Abhimanyu Chowdhary
Hello evryone, I am useing net-snmp-5.3.3 on MIPS architecture. I am getting following error messages in snmpd.log during initial snmpd bootup: netsnmp_assert !"registration != duplicate" failed agent_registry.c:536 netsnmp_ subtree_load() netsnmp_assert !"registration != duplicate" failed agent_re

RE: Reference material for creating entirely dynamic subagents?

2010-08-23 Thread Marc Butler
> From: [email protected] [mailto:[email protected]] On > subagents? > > On 22 July 2010 15:42, Marc Butler wrote: > > I am trying to implement an intermediate interface that facilitates > an > > entirely dynamic (no mib) subagent > > Can you explain a bit what you mean by this

Re: configure: This platform does not yet support hrSWInstalledTable rewrites

2010-08-23 Thread Wes Hardaker
> On Mon, 23 Aug 2010 16:53:54 -0400, Robert Story > said: RS> I was able to build a mini-agent (no swinst) and default agent (with swinst) RS> with this patch. The only thing I don't like is that the mini-agent is still RS> linking with the rpmlibs, even though they aren't needed. We r

Re: CFV: fix some tls tests

2010-08-23 Thread Thomas Anders
Robert Story wrote: > do you think a CFV is needed for changes to non-default test scripts? i.e. > scripts that won't be run unless explicitly requested by the user? IMHO no. In case others disagree, consider this a +1 for your patch. :-) +Thomas

Re: configure: This platform does not yet support hrSWInstalledTable rewrites

2010-08-23 Thread Robert Story
On Thu, 22 Jul 2010 10:56:22 +0200 Niels wrote: NB> Given that I am running RedHat 5 and Ubuntu 10.04 I am a bit surprised NB> to see (keep seing) this message. yeah, I've been seeing it on Fedora 12/13 too. NB> It seems that this is caused by defining HAVE_LIBRPM depending on it NB> being requir

CFV: fix some tls tests

2010-08-23 Thread Robert Story
First, a poll: do you think a CFV is needed for changes to non-default test scripts? i.e. scripts that won't be run unless explicitly requested by the user? Now the patch: this patch: - makes a cert name unique, since the cert util fingerprint lookup apparently does a prefix match, resulting i

Re: RFV: Do not use opensll libs unless HAVE_LIBCRYPTO is defined

2010-08-23 Thread Robert Story
On Mon, 23 Aug 2010 17:04:47 +0100 Dave wrote: DS> If you check the code, the only place where the two variables DS> actually appear in this file *is* within just such an #ifdef block. well then, +1 -- This SF.net email

Re: RFV: Do not use opensll libs unless HAVE_LIBCRYPTO is defined

2010-08-23 Thread Wes Hardaker
> On Mon, 23 Aug 2010 17:04:47 +0100, Dave Shield > said: DS> If you check the code, the only place where the two variables DS> actually appear in this file *is* within just such an #ifdef block. DS> I'm not sure why - since this functionality doesn't seem to be DS> tied in to the pres

Re: RFV: Do not use opensll libs unless HAVE_LIBCRYPTO is defined

2010-08-23 Thread Dave Shield
On 23 August 2010 16:54, Wes Hardaker wrote: > RS> updating existing ifdefs seems fine... what's with the extra ifdefs for > the > RS> variable decls? I'd expect to see a corresponding ifdef where the vars are > RS> used. > > I agree with Robert: I questioned the exact same thing. If you check

Re: RFV: Do not use opensll libs unless HAVE_LIBCRYPTO is defined

2010-08-23 Thread Wes Hardaker
> On Mon, 23 Aug 2010 09:55:32 -0400, Robert Story > said: RS> updating existing ifdefs seems fine... what's with the extra ifdefs for the RS> variable decls? I'd expect to see a corresponding ifdef where the vars are RS> used. I agree with Robert: I questioned the exact same thing. --

Re: RFV: Do not use opensll libs unless HAVE_LIBCRYPTO is defined

2010-08-23 Thread Robert Story
On Mon, 23 Aug 2010 07:10:00 +0200 Magnus wrote: MF> At the moment --without-openssl fails to prevent the use of openssl in MF> snmpusm.c. MF> The test case is an x86_64 host with 64-bit openssl libs but not 32-bit MF> openssl libs and you build a 32-bit net-snmp. MF> MF> The attached patch fixes

Re: RFV: Do not use opensll libs unless HAVE_LIBCRYPTO is defined

2010-08-23 Thread Bart Van Assche
On Mon, Aug 23, 2010 at 7:10 AM, Magnus Fromreide wrote: > At the moment --without-openssl fails to prevent the use of openssl in > snmpusm.c. > The test case is an x86_64 host with 64-bit openssl libs but not 32-bit > openssl libs and you build a 32-bit net-snmp. > > The attached patch fixes that

RE: TRAPS sent for excluded MIBS

2010-08-23 Thread mayank.mittal1
Thanks Dave, for the promp reply. Will definitely share the fix once available. Please do suggest on this last Question: Do we have snmp v3 trap supported in this version? The reason being we have two open defects which are as follows : 1)SNMP v3 user configured with AES privacy , when queried wi

Re: TRAPS sent for excluded MIBS

2010-08-23 Thread Dave Shield
On 23 August 2010 08:19, wrote: > Do we have any patches now to support this notify view over traps ? No. Nobody has submitted anything to address this issue since the discussion earlier this year. > Please suggest as this has become a high priority defect in our code. > Need your valuable gui

TRAPS sent for excluded MIBS

2010-08-23 Thread mayank.mittal1
Hi Dave, I am using net-snmp-5.4.1.2. Do we have any patches now to support this notify view over traps ? Please suggest as this has become a high priority defect in our code. Need your valuable guidance. Thanks, Mayank On 17 February 2010 12:09, Mayank Mittal mailto:[email protected]>> wro