As far as I know, you have to implement the queue yourself. The solution I
used for traps generated by my own code was to append traps, in the form
of snmptrap commands, to a script that I could execute later when the
interface
came up.
The problem is identifying when you should actually send
. That is, I gave the configure
script --with-cflags=-1 -O1.
So the crashes seem to be related to using the legacy g77 fortan
compiler and C optimization level 2.
Cheers.
On 5/9/11 9:16 AM, John Hardin wrote:
I'm experiencing a segfault when I use the override command
in the snmpd.conf file to set
It simply means that the library (and hence the agent)
will be able to recognise a fuller list of MIB object names,
Dave
Sorry to butt in, but how does this affect the agent?When does the
agent even see a human-readable name from a MIB file? I thought
it would only see the encoded
We've been trying to get Net-SNMP 5.4 to build without
embedded Perl support, but keep getting errors. Our toolsmith
found this in the configure.in file in the PERL SUPPORT
section:
if test x$install_perl = xno ; then
install_perl=try
fi
It appears not to take no for an answer.
Thomas Anders wrote:
John Hardin wrote:
We've been trying to get Net-SNMP 5.4 to build without
embedded Perl support, but keep getting errors.
What configure options are you using?
./configure --disable-embedded-perl --without-perl-modules ...
should work for you.
+Thomas
Mbps interfaces (on Linux 2.4) were all reporting 10 Mbps.
Does the interface command in snmpd.conf still work for the
case where the new code doesn't successfully detect the speed?
John
Dave Shield wrote:
On 29/11/2007, John Hardin [EMAIL PROTECTED] wrote:
Last I looked, Net-SNMP just
Last I looked, Net-SNMP just guessed at ethernet interface speeds. It
used to always guess 10 Mbps, but at some point it started to guess 100
Mbps. It does not dynamically detect the speed. You can change what
the agent reports by telling the agent the interface speed using the
interface
Thomas Anders wrote:
Scott Walker wrote:
It looks like after walking the entire squid OID, it kills snmpd.
This rings a bell. Can you please retry with 5.2.3.rc1 or 5.3.1.rc2?
We've fixed at least one issue wrt. squid proxying.
+Thomas
Take a look at bug 1501258 (5.3.1.pre3
Mike Lindsey wrote:
I've got an issue with snmp proxies. I'm using net-snmp 5.0.9... in
my snmpd.conf I have:
proxy -v 1 -c public localhost:3401 .1.3.6.1.4.1.3495.1
I can do an: snmpwalk -v 1 -c public localhost:3401 .1.3.6.1.4.1.3495.1
and I get the expected data (squid snmp output).
I am getting snmpd cores each time I walk the Squid MIB, which I
am proxying with the Net-SNMP agent using the following line
in snmpd.conf.
proxy -v 1 -c public 127.0.0.1:3401 .1.3.6.1.4.1.3495
I have recompiled the agent to add -g to the cflags so you can see the
calling parameters in the
Thomas Anders wrote:
John Hardin wrote:
I am getting snmpd cores each time I walk the Squid MIB, which I
am proxying with the Net-SNMP agent using the following line
in snmpd.conf.
Thanks for the report. Please file a bug on http://www.net-snmp.org/bugs
so it won't get lost.
+Thomas
[EMAIL PROTECTED] wrote:
Hi,
I am using Net-SNMP 5.3.0.1
I have compiled it with disman event mib and host mib support and
applied the following official patches.
1473289 5.3.0.1 SIGHUP breaks daemons
1429059 5.3.0.1: DISMAN monitoring crash
1421725 5.3.0.x: OID lookups
Thomas Anders wrote:
John Hardin wrote:
I have a fully functional agent that appears to forget about all
objects whenever it gets a SIGHUP. I have been using SIGHUP
to cause the agent to re-read snmpd.conf after a config change.
This has been working since UCD-SNMP 4.2 up through Net-SNMP
Thomas Anders wrote:
John Hardin wrote:
$ /usr/sbin/snmpd -c /etc/snmp/snmpd.conf -C 127.0.0.1:161
10.1.0.36:161
Don't do. You'll lose your persistent configuration on a restart. I'd
rather recommend:
env SNMPCONFPATH=/etc/snmp:/var/net-snmp /usr/sbin/snmpd 127.0.0.1:161
10.1.0.36:161
I have a fully functional agent that appears to forget about all
objects whenever it gets a SIGHUP. I have been using SIGHUP
to cause the agent to re-read snmpd.conf after a config change.
This has been working since UCD-SNMP 4.2 up through Net-SNMP
5.2.1. I recently upgraded to 5.3.0.1 and now
I've been having some problems with mib2c generating the
right code in the old ucd-style. I'm using NET-SNMP 5.2.1
on Linux. It's pretty easy to tweak the resulting code to fix the
problems, but I thought I'm mention the problems in case I'm
doing something obviously wrong.
The two problems
If you don't need the agent to send any other traps, you can just
remove the trapsink/informsink links from the configuration.
Then send your traps yourself.
pingping.han wrote:
Yes. It must hack the codes.
You can modify the Agent/snmpd.c file:
Find the following lines:
/*
* Send
Dave Shield wrote:
On Wed, 2005-10-05 at 12:48 -0700, John Hardin wrote:
I don't know how widespread this is, but I've dealt with a customer who
wants all traps to report the entrance or exit of a state.
Unfortunately not all traps fall into this model - whether your
customer likes
Robert Story wrote:
On Wed, 8 Jun 2005 09:39:24 -0400 Robert wrote:
RS That does sound like a bug in our agent. It should move on to the next
RS tree. I'm not familair with the proxy code, but I can take a quick look and
RS see if it's something easy.
I'm not sure if this will fix the issue,
Robert Story wrote:
On Wed, 8 Jun 2005 09:39:24 -0400 Robert wrote:
RS That does sound like a bug in our agent. It should move on to the next
RS tree. I'm not familair with the proxy code, but I can take a quick look and
RS see if it's something easy.
I'm not sure if this will fix the issue,
Robert Story wrote:
On Wed, 08 Jun 2005 09:36:07 -0700 John wrote:
JH I'll give it a try, but I got a fix from the Squid folks that's been
JH working for me. See bug #1317 in the Squid Bugzilla database.
JH The Net-SNMP agent receives the getnext for the last OID returned
JH by my code and
Thomas Anders wrote:
John Hardin wrote:
I'm using NET-SNMP 5.2.1 which I have extended with my own
MIB and configured to proxy for Squid (the caching software).
Squid is running on the same machine but implements its own agent,
which is listening on UDP port 3401. The problem is that when I
John Hardin wrote:
Thomas Anders wrote:
John Hardin wrote:
I'm using NET-SNMP 5.2.1 which I have extended with my own
MIB and configured to proxy for Squid (the caching software).
Squid is running on the same machine but implements its own agent,
which is listening on UDP port 3401
John Hardin wrote:
John Hardin wrote:
Thomas Anders wrote:
John Hardin wrote:
I'm using NET-SNMP 5.2.1 which I have extended with my own
MIB and configured to proxy for Squid (the caching software).
Squid is running on the same machine but implements its own agent,
which is listening
Wes Hardaker wrote:
On Mon, 02 May 2005 16:44:37 -0700, John Hardin [EMAIL PROTECTED] said:
John trace: init_agent_snmp_session(): snmp_agent.c, 1244:
John snmp_agent: agent_sesion 0x80f18c0 created
John Segmentation fault
Can you run it in gdb and send us the results of a where call
John Hardin wrote:
I have a system where I set up Net-SNMP to proxy for Squid
(the caching software). Squid is running on the same machine
as Net-SNMP. Its agent is integrated into its cache daemon
and only understands SNMPv1, so I have Squid's agent
listening on 127.0.0.1:3401.
The snmpd.conf
Wes Hardaker wrote:
On Mon, 02 May 2005 13:12:47 -0700, John Hardin [EMAIL PROTECTED] said:
John Apparently the BULK-GET received by snmpd is sent to Squid as
John a GET-NEXT with two OIDs. Ethereal reports this:
John Version: 1 (0)
John Community: public
John PDU type: GET-NEXT (1
Wes Hardaker wrote:
On Mon, 02 May 2005 13:12:47 -0700, John Hardin [EMAIL PROTECTED] said:
John It looks like the agent then segfaults when the snmpbulkget retries.
This code should have been executed in that case. Did you see this
message in your log:
if (request || var
John Hardin wrote:
Wes Hardaker wrote:
On Mon, 02 May 2005 13:12:47 -0700, John Hardin
[EMAIL PROTECTED] said:
John It looks like the agent then segfaults when the snmpbulkget
retries.
This code should have been executed in that case. Did you see this
message in your log
I have a system where I set up Net-SNMP to proxy for Squid
(the caching software). Squid is running on the same machine
as Net-SNMP. Its agent is integrated into its cache daemon
and only understands SNMPv1, so I have Squid's agent
listening on 127.0.0.1:3401.
The snmpd.conf file I give to the
Robert Story wrote:
On Thu, 17 Feb 2005 14:47:12 -0800 John wrote:
JH Arggh... Problem solved. Further research showed enterprises.3495 is where
JH the Squid MIB is registered. The agent was proxying for Squid, but
JH Squid was down. That broke the walk and caused the error
That shouldn't
Robert Story wrote:
On Thu, 17 Feb 2005 14:47:12 -0800 John wrote:
JH Arggh... Problem solved. Further research showed enterprises.3495 is where
JH the Squid MIB is registered. The agent was proxying for Squid, but
JH Squid was down. That broke the walk and caused the error
That shouldn't
Robert Story wrote:
On Wed, 16 Feb 2005 18:02:42 -0800 John wrote:
JH I've implemented a new MIB and am having trouble walking
JH some of the tables. [...]
JH I've traced this with Ethereal and found that when a GET-NEXT
JH is received for the last row of the last column of a table and I can
JH
John Hardin wrote:
Robert Story wrote:
On Wed, 16 Feb 2005 18:02:42 -0800 John wrote:
JH I've implemented a new MIB and am having trouble walking
JH some of the tables. [...]
JH I've traced this with Ethereal and found that when a GET-NEXT
JH is received for the last row of the last column
I've implemented a new MIB and am having trouble walking
some of the tables. After successfully walking the whole table,
the client reports an error or timeout. This new MIB is broken up
into multiple source files, each implementing a portion of the
MIB and registering its branch of the MIB
Mike Mohamet wrote:
Hi all,
I updated my agent from ucd-snmp-4.2.6 to net-snmp-5.2 and although
all the MIBs I had created works ok, the agent seems not to send any
traps (Cold start, Authentication failure etc.).
I was wondering if there are any further addition I have to make in
order to
Darren Gamble wrote:
Good day,
Hmmm... I decided to try this myself since I'm also running
Fedora Core 2.
I'm seeing the same results.
I ran
configure
make
make install
I configured two trap2sink targets, one to localhost and the
other to a
remote system. When I start and stop
Some questions:
Did a previous version of Net-SNMP work on this same system and
stop working when you upgraded to 5.2, or is this a whole new system
(or one where you upgraded the OS and Net-SNMP at the same time)?
When you say mostly works does that mean most traps get to their
intended
I don't really disagree with this, but would like to mention
that comercial vendors decide whether to fix a problem based
on their own commercial interests. Somthing critical for you
may not be critical for them. With open source you always
have the option of fixing a problem yourself or hiring
Hmmm... I decided to try this myself since I'm also running Fedora Core 2.
I'm seeing the same results.
I ran
configure
make
make install
I configured two trap2sink targets, one to localhost and the other to a
remote system. When I start and stop snmpd, the trap is delivered locally
but
...but make sure you haven't specified values for these objects
in your snmpd.conf file, or they will be treated as read-only.
Jeff Johnson wrote:
sysDescr is a read-only MIB object. try sysName or sysLocation or
sysContact...
/jeff
-Original Message-
From: [EMAIL PROTECTED]
How about counting the number of entryies in the Host Resources
hrProcessorTable. That works on Linux.
Bruce Shaw wrote:
I don't think that particular bit of code has been implemented on linux.
I'm not sure about HP/UX. It does work on Solaris.
Patches welcomed.
Can anybody please let me know
I believe HUP just tells the agent to re-process it's configuration
file, not restart it.
Marco Waller wrote:
Hi,
'kill -HUP' restarts the agent, but without resetting sysUpTime.
Is there another signal to do this ? Or do I have to kill and
start the agent again ?
Thanks, Marco.
43 matches
Mail list logo