rpki-client: Inconsistent RPKI repository caching results

[fjahr]

We are sometimes getting more RPKI repositories than on other runs. IRR always completes the data but it would be good to understand what is causing this and if there is anything that we can improve on our side.

I opened an issue in the rpki-client github: rpki-client/rpki-client-portable#116

[job]

It is best to use the same repository across multiple runs, not start empty every time

[fjahr]

Ok, I think we can try to add something like this for our scheduled runs. We can try to sync once before the we actually run and then just resync with the same cache dir when our run actually starts.

[job]

I recommend to somehow sync at least once an hour, this increases the chance to overcome intermittent transient reachability issues repositories might have