This is a discussion on Re: 5.4.1 won't talk within the SNMP Coders forums, part of the Networking and Network Related category; --===============0136605734== Content-Type: multipart/alternative; boundary="----=_Part_6413_7829689.1208356770696" ------=_Part_6413_7829689.1208356770696 Content-Type: text/plain; charset=ISO-8859-1 ...
|
|||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
|
|||
|
--===============0136605734==
Content-Type: multipart/alternative; boundary="----=_Part_6413_7829689.1208356770696" ------=_Part_6413_7829689.1208356770696 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline Good call, Dave: I put the line in snmpd.conf and restarted. Everything came up happily (and I noticed that /var/agentx/master exists as a socket, where even the directory /var/agentx had not existed before). For some reason it is behaving as if the subagent is not there, i.e. skipping over our subtree when I do a "getnext". But I expect I can turn on DEBUG in the subagent and track it down. If not, I'll get back to you... Thanks, Larry On 4/16/08, Dave Shield <D.T.Shield@liverpool.ac.uk> wrote: > > On 15/04/2008, Larry Dickson <ldickson@cuttedge.com> wrote: > > netsnmp_assert !"registration != duplicate" failed > > agent_registry.c:535 netsnmp_subtree_load() > > netsnmp_assert !"registration != duplicate" failed > > agent_registry.c:535 netsnmp_subtree_load() > > netsnmp_assert !"registration != duplicate" failed > > agent_registry.c:535 netsnmp_subtree_load() > > You can ignore those - they're harmless. > > > > Turning on AgentX master support. > > Error: Couldn't open a master agentx socket to listen on (): > > Who are you running this agent as? Root, or a different user? > > Do you have an "agentXSocket" directive in your snmpd.conf file? > What AgentX-related config settings do you have? > > Does the directory /var/agentx exist before you start the agent? > What about the socket /var/agentx/master > (both before and after starting the agent). > > > > This means agentx_sockets = > > strdup(netsnmp_get_string(NETSNMP_DS_APPLICATION_I D, > > NETSNMP_DS_AGENT_X_SOCKET)) returned nothing > > That's normal. > This is the mechanism for overriding the default AgentX socket > (either via the 'AgentXSocket' config directive, or the -x command > line option). > > > The issue actually lies with setting the default value (a few lines > later). > This changed from > agentx_sockets = strdup( {NETSNMP_}AGENTX_SOCKET ); > > in the 5.2.x and 5.3.x lines, and is now > agentx_sockets = strdup( "" ); > > This was part of SVN revision 15258, aimed at introducing a > more flexible transport creation mechanism. It may be that > this has broken (or changed) the default AgentX behaviour. > > A simple workaround would be to specify the AgentX port > explicitly - either in the snmpd.conf file > > agentXSocket /var/agentx/master > > or on the command line > > snmpd -x /var/agentx/master ..... > > Dave > ------=_Part_6413_7829689.1208356770696 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline <div>Good call, Dave: I put the line in snmpd.conf and restarted. Everything came up happily (and I noticed that /var/agentx/master exists as a socket, where even the directory /var/agentx had not existed before). For some reason it is behaving as if the subagent is not there, i.e. skipping over our subtree when I do a "getnext". But I expect I can turn on DEBUG in the subagent and track it down. If not, I'll get back to you...</div> <div> </div> <div>Thanks,</div> <div> </div> <div>Larry<br> </div> <div><span class="gmail_quote">On 4/16/08, <b class="gmail_sendername">Dave Shield</b> <<a href="mailto:D.T.Shield@liverpool.ac.uk">D.T.Shiel d@liverpool.ac.uk</a>> wrote:</span> <blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">On 15/04/2008, Larry Dickson <<a href="mailto:ldickson@cuttedge.com">ldickson@cutte dge.com</a>> wrote:<br> > netsnmp_assert !"registration != duplicate" failed<br>> agent_registry.c:535 netsnmp_subtree_load()<br>> netsnmp_assert !"registration != duplicate" failed<br>> agent_registry.c:535 netsnmp_subtree_load()<br> > netsnmp_assert !"registration != duplicate" failed<br>> agent_registry.c:535 netsnmp_subtree_load()<br><br>You can ignore those - they're harmless.<br><br><br>> Turning on AgentX master support.<br> > Error: Couldn't open a master agentx socket to listen on ():<br><br>Who are you running this agent as? Root, or a different user?<br><br>Do you have an "agentXSocket" directive in your snmpd.conf file?<br> What AgentX-related config settings do you have?<br><br>Does the directory /var/agentx exist before you start the agent?<br>What about the socket /var/agentx/master<br>(both before and after starting the agent).<br><br><br> > &nbs p; Th is means agentx_sockets =<br>> strdup(netsnmp_get_string(NETSNMP_DS_APPLICATION_I D,<br>> NETSNMP_DS_AGENT_X_SOCKET)) returned nothing<br><br>That's normal.<br>This is the mechanism for overriding the default AgentX socket<br> (either via the 'AgentXSocket' config directive, or the -x command<br>line option).<br><br><br>The issue actually lies with setting the default value (a few lines later).<br>This changed from<br> agentx_sockets = strdup( {NETSNMP_}AGENTX_SOCKET );<br> <br>in the 5.2.x and 5.3.x lines, and is now<br> agentx_sockets = strdup( "" );<br><br>This was part of SVN revision 15258, aimed at introducing a<br>more flexible transport creation mechanism. It may be that<br> this has broken (or changed) the default AgentX behaviour.<br><br>A simple workaround would be to specify the AgentX port<br>explicitly - either in the snmpd.conf file<br><br> agentXSocket /var/agentx/master<br><br>or on the command line<br> <br> snmpd -x /var/agentx/master .....<br><br>Dave<br></blockquote></div><br> ------=_Part_6413_7829689.1208356770696-- --===============0136605734== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757...un.com/javaone --===============0136605734== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Net-snmp-coders mailing list Net-snmp-coders@lists.sourceforge.net https://lists.sourceforge.net/lists/...et-snmp-coders --===============0136605734==-- |