Periodically push the netcfg instead of one time push

[hwchiu]

• Change the container image, we need jq and curl now
• Check the device value every 30 seconds

[teone]

@hwchiu @charlesmcchan we have a similar thing in VOLTHA, you have to be especially careful regarding the cfg for specific apps, as if the app is not active it will override the config with the defaults once it starts.

This was our solution: https://github.com/opencord/voltha-helm-charts/blob/master/voltha-infra/templates/onos-config-loader-script.yaml#L24-L89 it would probably be good to find a common solution to the issue and include it in the onos-classic helm chart. What do you think?

[charlesmcchan]

@hwchiu Teo actually brought up a good point - it is better to wait until ONOS is fully ready before pushing the netcfg. We only do it in wait-onos initContainer but that is done only once. In the scenario we are trying to fix (entire k8s cluster down), the pod might not be restarted and therefore the initContainer won't be effective. It is better to move the wait-onos logic into the while/sleep loop.

[charlesmcchan]

@hwchiu Please see if you can consolidate our script with the one VOLTHA is using. It'll be nice if we can maintain one jointly moving forward.

[teone]

I can look into that as well, once we're done with the release (should be soon).

[hwchiu]

@teone Thanks for the sharing!
@charlesmcchan Got it, I think I can change the wait-onos first to ensure TOST works correctly, and then we can figure out how to move all scripts into the onos-classic so that all projects can benefit from it.