Then you may contact the team to add your satellite to the Community satellites list accordingly mentioned program for the satellite operators at Phase 1. We did not move further this phase yet.
There is no automatic rewarding, unless you figure out how to attract customers and implement your own integration with payment processors to accept payments from these customers and implement a rewarding of nodes joined to your satellite (you also need to attract Node Operators as well).
This is still actual:
I know. Thats my main problem right now.
Reverse-engineering is a very hard work, so I need some help from the developers.
Maybe a few of my questions can be answered?
What is the recommended database? storjup uses cockroach, the boostrap shell script uses postgresql. Since I have only experience with postgresql: Are there any storj-sattelite specific benefits to use cockroach instead? Any difference for mediainfo and master database?
The storjup splits the satellite into 3 main parts: API, Core and Admin. Can you describe their main tasks for each in 1-2 sentences? Is it required, to split them up into separate container executions or can they run in one instance?
It looks like there are dependencies between the 3 parts: Core and Admin needs the API. Is there also a dependency between the Core and Admin instance?
STORJ_ORDERS_ENCRYPTION_KEYS: What is this encryption keys used for? Is there any generator? If not, what is the format: Alpha-Decimal, Hex, or something else?
What is the preferred docker image to use? I know, that there exists two of them for more or less the same thing. But why? Which one is Storj.io using? The images on img.dev.storj.io sound more like for a testing / quick start. Is the registry on dockerhub the “serious” one?
What I mean with the original post: I don’t run just simply storjup or storj-sim. I need to understand the parts and its details to run them in a proper way. Otherwise, I can never move them properly into a K8s cluster environment.
After this works, I will join the Community Satellite program (to hopefully find some more nodes to do a wider test).
A few answers would help me and also other people to operate successfully a satellite.
it’s optimized to use a cockroach database, it works better in a distributed mode and allows a horizontal scaling. In the one node setup PostgreSQL can be more convenient, however it’s up on you. For the production load the cockroach database is better.
I think you may use any of them or build your own using a Makefile from the repository. Our CI regularly publish images to the docker registry, so you may use them too. The only difficulty is to use a matching versions of all components, thus images from img.dev.storj.io are more consistent, but they more for the testing/develop purpose.
I (personally) would recommend using Cockroach as they are used in production almost everywhere. Cockroach is postgresql compatible, so easy to switch. Also has good web UI to spot slow SQL queries.
We have some cockroach specific optimization (mainly using AS OF SYSTEMTIME) and some codepath maybe slower with Postgres (where we had cockroach only optimization).
But both are tested with unit tests.
Note: we use Postgres for some of the queue-s, as Cockroach hasn’t the best performance for temporary queue-like data. But you can start with cockroach as you can switch any time later if you have higher loads.
We use this one in production, but we add our branding HTML + CSS set on top of these images (but that’s private, as it includes our branding stuff)
You can use the public images from Docker hub. Should work well.
But I don’t see any serious risk in using the img.dev.storj.io images. We use them for integration testing, so they should work. But they are not released with the release process, so usually they are not the latest one…