Open holdenrehg opened 8 years ago
I'm setting up OPM for the first time and ran into an issue with the dispatcher dumping logs from perf_data/ into the database. The dispatcher is processing logs out of perf_data/ and logs are being transferred over to perf_data/ correctly.
I'm not sure I fully understand the situation. IIC, nagios does process perfdata and writes them in a perf_data/ directory. Then does nagios_dispatcher process this directory? If yes, are the perfdata files removed? Are there any lines in the wh_nagios.hub table in the opm database?
On another note, are there any setup scripts created for OPM? I would be extremely beneficial to > have a script for different systems to install all of the necessary pieces. The current install process > is pretty cumbersome compared to similar software.
Yes, this is unfortunately true. The main issue is (in my opinion) the Nagios configuration. Debian and RHEL packages have different configuration (config file location, spool directory and so on), so it's not really possible to provide a template to copy "as it".
I worked on a docker template some times ago, I can do some cleanup and publish if it's of any help?
The nagios_dispatcher does properly process that directory, the files are removed, but then no lines are created in the wh_nagios.hub table. Does the nagios_dispatcher log anywhere?
Could a startup script could have parameters passed in to point to the config location, spool location, or even like a yaml / json where you could define all system specific parameters to build from. I ask because I plan on using this on quite a few different systems and will be setting it probably a couple times a month. If I get to building a script then I will definitely make a pull request.
The nagios_dispatcher does properly process that directory, the files are removed, but then no lines > are created in the wh_nagios.hub table. Does the nagios_dispatcher log anywhere?
If "syslog=1" in the .conf file, it's send on syslog, otherwise it's on stdout and stderr. You can add "debug=1" to see if more information is displayed, and launch it without the --daemon (or daemon=0 in the .conf file)
Could a startup script could have parameters passed in to point to the config location, spool location, or even like a yaml / json where you could define all system specific parameters to build from. I ask because I plan on using this on quite a few different systems and will be setting it probably a couple times a month. If I get to building a script then I will definitely make a pull request.
I'll try to work on it. I fear that it'll ease only a small part of the install, but it's a first step :)
I'm setting up OPM for the first time and ran into an issue with the dispatcher dumping logs from perf_data/ into the database. The dispatcher is processing logs out of perf_data/ and logs are being transferred over to perf_data/ correctly.
The postgres log are not currently displaying errors any errors.
On another note, are there any setup scripts created for OPM? I would be extremely beneficial to have a script for different systems to install all of the necessary pieces. The current install process is pretty cumbersome compared to similar software.
Holden