NetXMS Support Forum

Ok now im confused       i   just built 2.1 from source and now things went from --with-mqtt  to --with-mosquitto   there is no documention on mqtt and at all on the website  can some one throw me a hint on how to actually get mqtt working  and what the architecture is   does netxms now act as a broker and the agents publish  metrics to netxms  and will a third party software or dashboard be able to subscribe to those metrics to show which agents/nodes are up or down with out having to login using netxms client


Thanks

Victor,


What metrics are available for the agent to publish ? And where does the netxms server come into play ? from the looks of it the agent publishes messages direct to a broker  so then does the netxms server subscribe to that broker as well or is the netxms server acting as the broker ?   

Just curious.. I see experimental mqtt support was added and looking at the source it is only listed under subagents.  Is both the broker and client running on the agent ?   Just trying to figure out the topology of this. Does the agent publish predefined metrics to the server ?



Thanks,

knight,

I was able to get it working by compiling the 2.1-M2 and enabling Subagent =  ssh.nsm  and in the ssh settings add the credentials of the switch as well as the select the agent acting as a proxy. In the parameters all i did was specify the command I wanted to run on remote switch. See the attachments   I used a different machine to act as the proxy but I do think you can do all this from the server if you enable the agent locally and choose the local agent as the proxy.  One thing I did notice is that if the response is multi line you wont get the full response .. Im hoping the dev's can chime in and maybe provide some insight ...The custom commands on this switch a returned in json format so i only get beck the first line.

maybe check this thread it may have your answer


https://www.netxms.org/forum/general-support/dci-checks-over-ssh/

Never mind I got it fixed... I manually removed all nextms libs in my /usr/local/lib  and ran make install again and it worked not sure what happend 

getting the following when trying to install..  this is on a Ubuntu 16.04 system installing from source 


/usr/local/bin/nxmibc: symbol lookup error: /usr/local/lib/libnxsnmp.so.2: undefined symbol: wfopen
Makefile:1204: recipe for target 'install-data-hook' failed
make[5]: *** [install-data-hook] Error 127
make[5]: Leaving directory '/home/nmsadmin/netxms-2.1-M2/contrib/mibs'
Makefile:1139: recipe for target 'install-data-am' failed
make[4]: *** [install-data-am] Error 2
make[4]: Leaving directory '/home/nmsadmin/netxms-2.1-M2/contrib/mibs'
Makefile:1094: recipe for target 'install-am' failed
make[3]: *** [install-am] Error 2
make[3]: Leaving directory '/home/nmsadmin/netxms-2.1-M2/contrib/mibs'
Makefile:431: recipe for target 'install-recursive' failed
make[2]: *** [install-recursive] Error 1
make[2]: Leaving directory '/home/nmsadmin/netxms-2.1-M2/contrib'
Makefile:481: recipe for target 'install-recursive' failed
make[1]: *** [install-recursive] Error 1
make[1]: Leaving directory '/home/nmsadmin/netxms-2.1-M2'
Makefile:775: recipe for target 'install' failed
make: *** [install] Error 2

I have several switches that Netxms  2.0.5 is communicating with SNMPV3   I have seen this issues more than once but never thought too much about it but now its becoming a pain.  For some reason after configuration probably after the configurations polls are executed netxms  sees these layer 2 switches as having agents   which breaks any DCI and summary tables these are associated with..  Any suggestions as to how I can stop this or is this just a bug



Thanks

why not just use syslog-ng to grab the fail2ban logs and ship them to netxms syslog

Is it possible to launch on demand topology polls and network discoveries via the command line ?

Is it possible to get snmp values via nxshell ?  I need to get the serial numbers for a certain type of switch and the only way is via snmp




Thanks

I havent  done too much with ERL other than use it to monitor the device itself but on the ERPro  I have a few hundred devices I use it to do snmp checks on and I havent really noticed any issues with performance at all



P.S
I dont have any x devices but if it is still based on debian like the ERL then it should work... Try at your own risk

Sure,

I normally build an agent for each new release since I normally upgrade anyway ...I will work on 2.0.6 tomorrow

Can we start a thread or a place for people to post scripts.  I think it would be useful..Im sure there are alot of people that have scripts that could be useful for new comers or some that maybe trying to do the same thing.





Thanks

Its working now ... Thanks for all the help