Community Satellite Spiridon is back

I don’t qualify for the 20 nodes but quite happy to donate space (I have plenty!) for this initiative.
Can I do that?

Because I need to think about the legal consequences first.

But if you are interested about such a scheme, let me know.

Same answer. I think it’s reasonable. But in this case I would feel it more fair to contribute some amount to the cost of maintenance.

You can join the satellite any time.

My current plan is:

  1. adopting the code from the original spiridon satellite to support telegram based node status notifications.
  2. to make it usable for everybody, I will change invite robot to make it possible to join with 1 node only (== invite link to login to UI)
  3. but by default you won’t have the projects to actually upload / download (just manage your own nodes).
  4. if you have at least 20 nodes (enough for using your own nodes with the configured EC parameters) → a project will become available to you to use it
  5. if enough nodes are joining (like hundreds of operators) → an other project will be available to everybody: using the shared pool. Enable to use space slightly less what you offer (~with the EC expansion factors).

At least these are my current ideas. Let me know what you think.

OK, I’ll add my nodes
Just so that the moron that is me understands, I just need to add

storage2.trust.sources: https://www.storj.io/dcs-satellites,storj://1NusSk8HjWppghiWvofEBDqryDaxiALPah8EyRxXWdNkwXg7ai@spiridon.space:7777

To config.yaml and I’ll add your satellite and all the other Storj ones will continue working as normal?

How about IP address? Would 20 nodes using same IP address be ok?

This will defeat the point. Why would you need to mess with storagenodes if everything is local - you can just write data to a NAS?

What point? This is a community satellite coming without any serious durability. Can’t see a benefit from using fake IPs here.

No, you got it backwards. Did you read the top post?

First – this is a pretty explicit requirement:

You will only be able to use space on your own nodes. If all your nodes are under the same roof – then just use a conventional nas to store your data, what do you need storj protocol for?

This community satellite only absolves you from running a satellite yourself. Nothing more.

Pleas read the top post.

If I had 20 real locations I would not use a community satellite. I would be surprised if there is any SNO having 20 real locations. :joy:

Maybe some small business with multiple offices. No idea. I too think it’s too niche.

“At least 2 weeks of vetting period”

So now I just have to wait for my nodes to be vetted (without any data to audit?), before I’m able to do anything?


Th3Van.dk

nodes cant be veeted without data, it makes audits.

I’m not able to create a bucket using the web interface, but using a grant access key and Uplink I’m was able to create a bucket, but i still can’t upload any data to it → uplink: segments limit exceeded

However, I had a lot of fun clicking “confirm” ownership 264 times manually :wink:

Th3Van.dk

This error usually occurs when some services are blocked, in this case most likely the authentication service—you can see this in the developer console, which is actually unresponsive.

Typically, to fix such errors, you need to unblock the service in your ad blocker/filtering software/hardware, or disable “smart” protection on your firewall/router/ISP, or elsewhere, or use the mobile browser or the mobile internet, in hard cases using VPN.

However, in this case, it’s also possible that the service is simply unavailable because it’s an experimental satellite.

ownership

Yeah, sorry. It’s not yet optimized for hundreds of nodes, yet.

The reason behind confirmation: as the main placement will store data on your own node, it might be possible that somebody joins to the satellite with faking your own email.

The user has a confirmed email, but you need to confirm your nodes.

the audit period

I don’t believe that audit period is necessary, if you use your own space. Also: it requires at least a few segments. That’s why I added the minimum 10TB reserved space, but likely I will ignore it.

The current behavior of the system:

  1. if new user joined (email) with at least 20 nodes, a new invite will be sent out. New project will be created with 0 limit.
  2. an other chore checks the conditions: if at least 20 nodes → a specific project will be created with the self hosted placement + placement will be created

The second part is still manual (but I am working on nice refactors to make it automatic).

Long story short, the current audit is something like this: I will create placement rules for new users at least once per week manually.

The placement is created for you @Th3Van

My original plan was to use usual rule:

  • EC parameters: 10 / 12 / 14 / 16
  • node selection: max one piece per node (that’s why I requested 20 nodes from different IPs)

But I am grateful for new testers to be joined, so I just set this placement to use fully random() because you have only one IP.

notifications

I also deployed the test version of the functionality of the old spiridon satellite: trigger node notifications via telegram. Sill testing, but should work.

other plans

I am considering to remove requirements to join to the satellite (20 nodes / 10 TB), as telegram notification can be useful for anybody.

But I would like to improve the project creation. By default only a project with 0 limits will be available, and all other specific projects will be auto-created when requirements met.

  1. project of a globally shared pool → only if enough operators are joining, with strict limits (use similar amount what you shared)
  2. pool of your own node → at least 20 nodes are required (because current EC parameters)

I am also considering to create project with test placement (1 replica), but I think it’s too confusing. It’s hard to communicate the unsafety.

Also considering to create shared pools (a few SNO which trust in each other can share space between the group). I am not sure about the UX experience, but if there is a demand, I can configure it manually for the time being.

I also have other ideas (like implementing a fully offline takeout plan → restore data from a bundle without satellite), but that is still in early phase.

I have no problems accessing EU1 satellite via the web interface, using the same browser.

Also tested with Chrome and Edge, but I get the same error :

I can’t upload using Uplink either…

Th3Van.dk

What I mean is that this satellite and its services are likely hosted on networks other than Storj’s. Blocklists operated globally, and there are so many “authoritative” providers of “legitimate” domains that you could spend a lifetime trying to find them all and convincing them that you pose no threat. And you still not a threat because of one of your user, who thinks otherwise. Welcome to the hell of consumers’ services…

It should be fixed by now. Can you please try again?

  • I made a the project provisioning smarter. It also changes the limits + adds descriptions. (I can add exemptions to test it without 20 nodes)
  • I re-deployed gateway with the right domain setting. UI based uploads are also working for me. I have to create an automatic test for this…

Thank you @elek - I’m able to create buckets, and list files now using the web interface :+1:

Uploads and downloads seem to be working also - but for some reason all uploads using Uplink are only going to 65.21.233.162 :

Please send me your PayPal - so i can donate for renting space, if that IP is Spiridon :wink:

Thank you for looking into this - much appreciated.

Th3Van.dk

That’s bad. That is my test placement.

Looks like the bucket inherited the wrong placement during early test phase.

I fixed the project placement previously, but bucket created earlier had old default placement.

Now it should be fixed. Unfortunately you should re-upload the data.

Or even better, you can create new bucket (just to make sure which data is replaced properly and which data is not)

Location is now visible in the UI and you should see: 5b3e564b521f4b11b59476346436c944 after every bucket.

(not very nice, I am thinking how can it be more user friendly (maybe email?), but this is how does it work now. I am kind of extending Storj to make all of these possible…

Thanks for the offer. Long term I am considering to open for donation, but right now, I am more focused to provide good service first. This is a small server, and I can afford it (at least while I have a job :wink: )

Bigger server may need a donation, but let’s make it working well, first.