I’m trying to deploy the Mysql Application Monitoring. I’m using the Agent method per this page:
https://docs.librenms.org/Extensions/Agent-Setup/
https://docs.librenms.org/Extensions/Applications/#mysql
I have followed both guides and ensured we have the proper port open in our Firewall. We can run the agent command just fine from each machine but the application pulls in no data. Any pointers on where to start troubleshooting this would be appreciated.
Agent script only works if root access to database is password protected and script has the correct password
The script runs locally just fine. We did provide the root password and the script has the correct password as we get output when we run it on the machine it’s installed on, we just get no data back from the Libre Pollers.
tried to run it localy under snmp Daemon user?
Our debain builds run it as Debian-snmp which is not allowed a shell login. When I run the command locally I get output but the files are owned by root. I’ve not found any logs either on the polling boxes or the servers I have the agent installed on to get any ideas there.
pro tip:
change in /etc/passwd TEMPORARLY on Debian-snmp /bin/false to /bin/bash
then change user to Debian-snmp with
su Debian-snmp
and now try to run mysql agent script
after testing restore settings in passwd!
1 Like
Thank you for that. I am able to run the script and get get output. I also see the agent-local-mysql file populate in /var/cache/librenms:
-rw-r–r-- 1 Debian-snmp Debian-snmp 1.5K Jun 16 09:56 agent-local-mysql
I can also telnet from my pollers (I have 2 of them) to each of the machines on TCP port 6556 and I get output from the file that is generated so firewall appears to be working good
In CentOS 7 didn’t have the luck with agent. Though terminal output was okay but didn’t get any data on web. However SNMP extend method worked just fine.
As I haven’t received any other feedback that’s probably what I’m going to end up doing. I forgot we tried to get the agent working for BIND as well and it wouldn’t work and we had to resort to the SNMP extend method there too.
If you are using the agent, the user that should be able to run the script is the agent’s one.