-5.2.1]#
I will next try a fresh download of the toolkit.
WRB
---
SF email is sponsored by - The IT Product Guide
Read honest candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start
incorporated into the
library libnetsnmpmibs.a, and this is the output from the nm function
[EMAIL PROTECTED] .libs]# nm libnetsnmpmibs.a | grep -i -n sysoid
77:0400 D version_sysoid
I do not get this error under toolkit version 5.2; it compiles under the
above configuration just fine.
WRB
Robert (and all the other core developers):
I have worked with the files
./agent/mib_modules.c
./agent/mibgroup/mib_module_inits.h
./agent/snmpd.c
to include printf() and DEBUGMSGTL(()) statements,
only to find, much to my chagrin, that those changes
that are made to the routines
should_init()
Robert (and the rest of the core developers):
I took the code of the routine init_mib_modules() and copied it to
the file snmpd.c, placing it just ahead of the function main(). I
actually kept it with the modified name of init_mib_modules_wrb().
Then made the daemon anew, and executed it. The
Robert:
I have some more information regarding this problem. First, I am using
RH 9.0, Net-SNMP 5.2.1 installed in the /root/net-snmp-5.2.1 directory,
with the mib2c-derived code for the 802.11 MIB in the directory
./agent/mibgroup/dot11/mfd (this time I have not deleted the .m2d files),
and am
Robert:
I have some more commentary on the MFD compilation. You will note that
this email is a follow-up to the most previous message, sent within the
previous two hours.
When I compile the daemon, using the configuration (noted here and in
the above mentioned previous posting)
./configure
, and
deposit it into the delete box, but do not forward it to others. All you
do by forwarding rejected messages is to further clog the bandwidth of the
internet.
WRB
Original message
Date: Sat, 29 Jan 2005 13:22:05 -0800
From: david [EMAIL PROTECTED]
Subject: Re: Undeliverable: WRB
Robert:
I have worked to get the dot11StationConfigTable linked into
the Net-SNMP daemon, with some interesting errors observed.
First, why is it that the headers files produced via MFD
mediated MIB compilation are not found during the build
process? Here is a relevant extract from build
I noticed the following segment in the function
init_snmp(const char * type)
as found in the file ./snmplib/snmp_api.c
/*
* make the type available everywhere else
*/
if (type !netsnmp_ds_get_string(NETSNMP_DS_LIBRARY_ID,
I have been looking at the files in
./agent/mibgroup/if-mib/data_access
and noted the content of the files
interface_ioctl.c and interface_linux.c
I take it that these are complementary files,
that they provide equivalent mechanisms versus
different operating systems in order to gain
On looking closely at the file
./agent/mibgroup/if-mib/data_access/interface_linux.c
we find the routine
netsnmp_arch_interface_get_if_speed()
and from what I can see of the implementation, it does
not support the determination of network interface
speeds in excess of 100MBit/Second.
Is
Robert:
WRB Wow. great description of your problem. I wish all our users were so
WRB verbose!
WRB
WRB It is really nice to have someone say this, how appreciative they are
WRB of the verbosity of my postings. Dave tends to complain about the
WRB length of same, suggesting that answers would
Robert:
WRB Since the table_dataset model operates through the
WRB retention of data internal to the agent, I would suspect that get requests
WRB will require that internal data store to be updated as appropriate prior
WRB to execution of the function
WRB
WRB netsnmp_table_data_build_result
WC I question the wisdom of requiring the function parameter to be
WC as specified. It seems more reasonable to pass *table_set* as
WC the parameter, instead of its element *default_row* being passed,
WC especially when the dereference can be easily accomplished from
WC within the function.
I just downloaded the latest release (net-snmp-5.2.tag.gz), and am
running on RedHat 9 (with current up2date support). The problem
with the MFD helper remains unresolved. The appropriate excerpt
from my command line is:
===
Excerpt Begin
Now, that is some facelift for the index page of the
Net-SNMP website. Generally, I like it. There are
two errors of which the webpage designer should be
made aware. They begin with the mention at the
bottom of the index page (alright, home page), where
the users mailing list is described.
Actually, I should have included that the incorrect
links are those for subscribing to the groups. The
links for getting to the archives are correct.
William R. Buckley
President
SoftNerd, A California Corporation
Director Emeritus,
International Core Wars Society
[EMAIL PROTECTED]
415-240-6107
Robert (and community):
All right, I'll bite. You say that the function
netsnmp_table_set_multi_add_default_row
specifies the default values of fields of a row, and that it does
not actually create a row. OK, OK. Yet, I still claim the naming
of the function is inappropriate. I would
Jeff:
I tried this construct:
netsnmp_table_row_add_index(row, ASN_INTEGER, 1, 0);
[...]
netsnmp_table_row_add_index(row, ASN_INTEGER, 2, 0);
and that failed to execute to completion (i.e., the agent
did not start - it did not show up as a process subsequent
to execution initiation, as measured
Why is it that the tutorial page, at URL
http://www.net-snmp.org/tutorial-5/agent/data__set_8c-example.html#a1
first shows how to define the default parameters for a
column of a row in a data_set table, and then uses the
function call sequence of
netsnmp_set_row_column(row, 2,
The relevant code is:
row = netsnmp_create_table_data_row();
netsnmp_table_row_add_index(row, ASN_INTEGER, 0, 0);
. At
the very least, they provide the option of alternative documentation, so
as to improve both my own understanding and mastery of the toolkit, and
for the similar improvement of other users of the toolkit.
WRB
PS - other questions given in my original posting on this topic have yet
to be addressed
The is also the Net-SNMP MIB. The relevant enterprise numbers are
UCDavis 2021
Net-SNMP 8072
Neither of these MIBs, UCD-SNMP and Net-SNMP, are specified via
IETF RFCs.
WRB
Original message
Date: Tue, 23 Nov 2004 17:39:49 -0500
From: Andy Smith [EMAIL PROTECTED]
Subject
the documentation for this function exists. Also, tell
me why the index cannot be *completed* as part of the operation of the
function
netsnmp_table_set_multi_add_default_row
I argue that the default row should mean the row indexed by 1, in this case.
WRB
Original message
Date: Tue, 23 Nov
This problem was apparently caused by the interpostioning of a firewall between
the host and target computers. This message should serve as a good source of
information regarding communication with an agent across some telecommunication
interface.
WRB
Original message
Date: Tue, 23
realnetgroupv2c realnetuser
viewfullviewincluded .1
access realnetgroup any noauth exact fullview fullview fullview
syslocation RealNet HQ
syscontact WRBuckley
rwuser wrb noauth
rouser brw noauth
rocommunity others
rwcommunity realnet
Some additional data may assist to understand the *walk* problem. From
the debug output comes:
===
Extract start
===
vacm:checkSubtree: , unknown
trace: snmp_call_callbacks(): callback.c, 231:
callback: END calling callbacks for maj=1 min=12 (1
As part of 802dot11 MIB implementation, there comes a point in the
mib2c generated code that tables are defined and constructed. Owing
to some perhaps unreasonable expectations, I am making changes to
the *netsnmp_table_set_multi_add_default_row()* function call, by
means of altering the
/mibs and exported
the environment variable MIBS=ALL. The first question asked by the script
is which api I would like to use. The extract from my command-line output
is:
===
Extract start
===
[EMAIL PROTECTED] wrb
Of course, the problem with mib2c in Net-SNMP 5.1.2 occurs only
in the *var-* routines, not in the *write-* routines.
William R. Buckley
President
SoftNerd, A California Corporation
Director Emeritus,
International Core Wars Society
[EMAIL PROTECTED]
415-240-6107
Well, I found another problem. As a good programmer should, I cleaned
up the problems found with the mib2c generated code for the 802dot11
MIB, and compiled to the point of generating the daemon, snmpd. At that
point, during linking, I obtained an error regarding an undefined reference
to a
defines?
Bill
Original message
Date: Fri, 19 Nov 2004 10:10:03 +
From: Dave Shield [EMAIL PROTECTED]
Subject: Re: WRB - mib2c anomaly.
To: [EMAIL PROTECTED]
Cc: NetSNMPUsers [EMAIL PROTECTED]
Well, now I have found more problems with mib2c in Net-SNMP 5.1.2
. The output from
listserve archives, for the benefit of
those who seek to extend the agent.
WRB
Original message
Date: Fri, 19 Nov 2004 10:18:19 -0500
From: Robert Story (Users) [EMAIL PROTECTED]
Subject: Re: WRB - Computing the next OID by an agent.
To: [EMAIL PROTECTED]
Cc: NetSNMPUsers [EMAIL
recognised the
benefit of my postings.
I have not as of yet searched other modules for the occurance of similar
*shutdown-* routines.
WRB
Original message
Date: Fri, 19 Nov 2004 14:05:15 -0500
From: Robert Story (Users) [EMAIL PROTECTED]
Subject: Re: WRB - mib2c anomaly.
To: [EMAIL
Here is a documentation error that I found while,
a) reviewing the code of the demo application at URL
http://www.net-snmp.org/tutorial-5/toolkit/demoapp/snmpdemoapp.c
b) reviewing the documentation page at URL
http://www.net-snmp.org/tutorial-5/toolkit/demoapp/index.html
The error
I am looking to understand the detailed process used by the
Net-SNMP toolkit of computing the next OID, such as is done
during a walk of a MIB. Some hints from the core developers
would be much appreciated, especially if same comes with
suggestions for code that I should review.
William R.
I am working with RH9, Net-SNMP 5.1.2, and the 802.11 MIB. I notice
that the output of mib2c is different, depending on use of the UCD
versus NET formats. Specifically, UCD generates names and OID's that
are not generated for NET. The particular case I find concerns the
OID named in the MIB as
several blank lines where that directive ought to be shown.
That is, the *figure* in the documentation is missing - it is blank.
WRB
Original message
Date: Wed, 17 Nov 2004 11:54:45 -0500
From: Robert Story (Users) [EMAIL PROTECTED]
Subject: Re: WRB: A seeming failure of mib2c in Net-SNMP
In reading the file AGENT.txt, I notice under
section 3 *The C code header file*, subsection
*Module dependancies* (yes, in the file the
word is misspelled), comes the first paragraph
suggesting the existence of a directive, yet
the place where the directive ought appear in
the file is blank.
I
OK, OK, I think I now see the problem correctly. It is
not with mib2c but with AGENT.txt, as the text file is
out of date. If I can get commentary on a rewrite, I
offer to provide same.
Are the core developers game?
William R. Buckley
President
SoftNerd, A California Corporation
Director
Sundeep:
MIBs are not solely for the purpose of informing a master agent
of the object manageable by a subagent. Indeed, MIBs to provide
this information but, not in the way that your questions suggests.
A MIB is an abstract description of data items and the structural
relationship of each data
Mauricio:
I recommend that you read the book Windows NT SNMP, and use its
examples as a means of first education regarding the use of SNMP
in a Windows environment. The author of the book has done a
superb job of making SNMP simple to understand, if it is a bit
abridged. The book is published
Alright:
It seems that my problem with the Kenati OS port of
Net-SNMP is that the daemon produced for Kenati is
not opening the UDP port (161). The Net-SNMP version
supplied with Kenati is 5.0.8, the daemon is running
on the target, I can walk the daemon on the target
from the target but, I
Moderators and developers:
Might it be possible for my gmail account to be
approved for direct posting to the newsgroup,
without my having to subscribe. After all, I do
not need two copies of postings, one to the
wrbuckley.com email, and one to the gmail account.
Contact me offline, should you
I replaced the 5.0.9 distribution with 5.1.2, and
observed the same problems in returning a UDP packet
for a get on MIB-II:SysDescr, as passed over ethernet.
On the other hand, I can fully walk the daemon from
the same computer as the daemon executes upon. So,
snmpwalk problem seems to be
/root/wrb/WRB/net-snmp-5.0.9
[EMAIL PROTECTED] net-snmp-5.0.9]# ls
acconfig.h config.cache configure-summary FAQ Makefile
mkinstalldirs README README.thread temp
aclocal.m4 config.guess COPYINGinclude Makefile.in
net-snmp-config
In making the port, I noticed a few things.
1) The file net-snmp-conf.h is constructed by the action
of the ./configure script. This is an important fact for
the Kenati port, as the definition of a defined constant
results, which causes the compilation to fail. The defined
constant include the
Robert:
Sure, I am happy to assist you with adding to the FAQ you offer.
In the next few days, I will forward the collection to you.
Perhaps you will be able to get he/she who maintains the Net-SNMP
FAQ to include within it a link to your FAQ. This would be
especially helpful if your link were
Robert:
It would have been helpful to me to have known of your FAQ when
I was working to port Net-SNMP to the Kenati embedded Linux
distribution. Perhaps you can extract from my postings in the
last two months and add the relevant information to your FAQ.
William R. Buckley
Original
Steve:
I think that your issue with the endianness of the target
processor, such as it prompted a bit of distress with
Robert Story, will perhaps come from not using a parameter
with the ./configure tool. That parameter specifies the
endianness of the target processor, and I suspect that use
of
Valley:
Do not post to the net-snmp-users list with html. Use only
text.
I suspect your answer is to be found in the FAQ. Failing that,
use the net-snmp-users list archive, found on the website
www.netsnmp.org under the menu item lists. If you use the
query function in the upper left hand
I am trying to port Net-SNMP into the Kenati Linux
environment. Kenati is an embedded environment, and
so the toolset is being used within a host RH9 Linux
distribution. As part of my effort, I am following
the process outlined for ./configure, a process that
is reported by Bob Ray (in his Aug
If the OID is defined within a standard MIB that is already
implemented via the Net-SNMP daemon (snmpd), then one expects
the set request to be rejected by the daemon. In fact, the
entire packet should be rejected, regardless of the number of
varbinds contained within the packet. If the OID is
I am trying to understand the make files for Net-SNMP,
so as to port the tool to an embedded Linux environment.
It is important that the build process not result in
an installation of the tool to the host development
environment, which is RH 9. It would be nice to avoid
modifying the existing
Additional observations are:
./agent Makefile is the correct make file for building snmpd
Modification of the compiler alias CC to proper directory
and executable yields a good attempt at building. Frankly, it
seems to me that the gcc compiler should be sufficient. Yet,
when the daemon is
55 matches
Mail list logo