Bad default location for REPORTS_ROOT

[bdlamprecht]

Current Behavior

Due to the way your current implementation references configuration files located in /etc/netbox/ instead of /opt/netbox/netbox/netbox/, when you browse to http://localhost/extras/reports/, the following is seen:

Yes, I know this can be changed in the netbox.env by setting a value for REPORTS_ROOT, however, in my opinion, this "default" is not a good idea. I don't believe anyone is going to use the /etc/reports location to store custom python scripts specific to NetBox.

I'd like to propose using /etc/netbox/reports/ as an alternative, but I didn't want to go through the hassle of creating a PR if it wasn't going to be merged.

What are your thoughts one way or the other?

Expected Behavior

N/A

Debug Information

N/A

[cimnine]

You're right, /etc/reports is a bad location for the reports. IMO this happened when we moved the 'main' configuration file to /etc/netbox and BASE_DIR now expands to /etc rather than /opt/netbox/netbox.

I agree that /etc/netbox/reports would be a good location. But I want to make sure this works out when using a configuration managed by Kubernetes/OpenShift. And I'm not entirely sure it does in a nice way.

IMO the nice way would be to have two separate ConfigMaps, one for reports and one for configuration. And when the one for configuration is mapped to /etc/netbox, I'm not sure if we can map the one for reports to /etc/netbox/reports.

So the main configuration might have to be moved to somewhere else, e.g. /etc/netbox/config, so that config and reports can co-exist.

Or am I over-complicating matters and it's fine to have one ConfigMap for config and reports?

I'd love to hear your opinion on this, and @tobiasge's as well, as he initially suggested to move the configuration to /etc/netbox.

[tobiasge]

Hello,

at the moment we're not using reports but I also think /etc/netbox/reports would be a good location for that. But you're right. You can't map to /etc/netbox/reports when the configuration is mapped to /etc/netbox in Openshift.
So the proposal with mapping to /etc/netbox/config and /etc/netbox/reports is a good solution.

I won't be able to implement this now because I'm on holiday for the next weeks. If this can wait I will do the implementation.

@flipkill1985 Could you test this in our deployment?

[cimnine]

@tobiasge Have you had the chance to look into this? Would one ConfigMap for config and reports be fine or do you think they should better be separated?

[tobiasge]

Hello,
I'm back now. I think it would be better to have them separated. This makes it possible for the user to also define a persistent volume in Openshift.