On Wed, Feb 9, 2011 at 7:52 PM, Dave Shield <[email protected]>wrote:
> How are you stopping the agent? Ctrl-C ?
>
I usually stop the agent with "Ctrl + \" . And I hate to tell you this, but
Ctrl-C is a bad way of stopping the agent. At least, with net-snmp 5.5,
whenever I used that, the agent would quit, but still leave locks on the UDP
port. At the time (a few months back), I asked around in the mailing list
and was told to use "Ctrl + \". That key combination has been working well
with me for stopping the agent without leaving UDP locks behind.
>> Any ideas why this is so? Could it be that the internal md5 and des
>> functions are not working?
>No - I don't think it's related to the internal security functions.
>This feels more like the handling of config files.
THIS reply of yours got me thinking - could it be something other than a
cross-compilation issue? So, I tried compiling & running the agent on my x86
Virutal system with the following config,
./configure --with-openssl=internal --enable-mini-agent --disable-shared
--disable-mibs --disable-applications --disable-manuals --disable-scripts
--disable-mib-loading --disable-embedded-perl --disable-perl-cc-checks
--without-rpm --with-out-mib-modules="ucd_snmp snmpv3mibs agent_mibs agentx
target disman/event disman/schedule notification-log-mib notification host
utilities" --with-mib-modules="utilities/execute"
To my surprise, I was able to re-create the error in my Virtual System as
well !! Now, what could be causing it? My first attempt had me running this,
sudo ./snmpd -f -Lo -DALL
which resulted in such a boatload of output that it was hardly any worth.
So, I modified it to,
sudo ./snmpd -f -Lo
This helped. One of the first log lines I got went something like - "unable
to process/identify token createUser". From there it was a matter of trail &
error that helped me find the problem - the "snmpv3mibs" !! If you notice, I
deliberately try to build my agent without it. The intention was to remove
the OID branch entry and not SNMPv3 authentication/security.
In retrospect, I can understand why a developer would skip building in the
crypto functions if snmpv3mibs were not present. Still, it sure was fun
trying to figure all this out! :)
Thanks a lot for your assistance Dave! It would have taken me another day or
two otherwise ...
-Abraham Varricatt
My technical side <http://blog.passion4software.com/>
------------------------------------------------------------------------------
The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE:
Pinpoint memory and threading errors before they happen.
Find and fix more than 250 security defects in the development cycle.
Locate bottlenecks in serial and parallel code that limit performance.
http://p.sf.net/sfu/intel-dev2devfeb
_______________________________________________
Net-snmp-coders mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
