Nagios Core4 and NConf

[thanstra]

I am working to upgrade both Nagios (up to 4.0.2) and NConf (up to 1.3.0) and something is not working properly in the new configuration.

I can get NConf up and running with my current database (upgraded) and can see all of my data, but when I attempt to deploy I get a very generic message:

Error: Cannot open main configuration file '/usr/local/nagios/nconf/temp/test/collector.cfg' for reading!

This message does not change no matter what configuration changes I try to make to either Nagios or NConf. It continues to look for this "main configuration file". I've tried rolling back NConf to 1.2.6 (with appropriate database rollback) but still get the exact same error message. I've looked all over for what wants "collector.cfg" but cannot find where this is coming from.

Does anyone have ideas of what is not fully configured? I'm assuming it is a config file somewhere using something generic but so far I've been unable to track it down.

Thanks,
Tom Hanstra

[slansing]

Have you verified the permissions are correct down that directory tree?:

Code: Select all

ls -la /usr/local/nagios/nconf/temp/test/collector.cfg
ls -la /usr/local/nagios/nconf/temp/test
ls -la /usr/local/nagios/nconf/temp

[thanstra]

There is no such file. I have no idea what is supposed to go into this file and cannot find any documentation on what this file is supposed to be.

It seems like NConf should be generating the configuration files, so I'm not at all understanding what kind of file it is attempting to *read*. Is there some default configuration file which is expected? What is this?

[thanstra]

To further follow up on this, I've been looking all over for an understanding of why NConf is looking for that particular directory structure and related files. In no place do I have anything like a temp/test/[file] configuration line, yet that is consistently and almost immediately where NConf is looking to read some sort of configuration file. If someone can explain why it is looking there, perhaps I'll be able to track this down further.

Thanks,
Tom

[sreinhardt]

I am not sure why it would be doing that, however I would suggest creating the requested folder structure and giving it 664 nagios:nagios access and seeing if something is created from nconf

[thanstra]

Tried that early on and it did not make a difference. If there is a file (even empty), it will at least run with no output. If there is only the directory structure with no file, it deletes the directories.

Very odd behavior compared to what I have seen earlier.

One thing that seems to be new with NConf is the deploy configuration. I did not use that in the past. Could that be involved?

[abrist]

Possibly, but you are in a rather untested/undertested country with nconf and nagios 4 as there have been multiple reports that nconf has compatibility issues with core 4. You may need to wait until nconf is up to date with the changes in core (specifically a few new directives).