How to setup Storj Windows GUI

Hello, I am trying to set up Storj Windows GUI using Virtualbox and creating multiple windows 10 installs.

My question is, are we allowed to create more then 1 Identity per person?

Yes! Of course! You need different Auth tokens though. You can use the same email to create different Auth tokens. I think you can only get one token per day with the same email though. If you are making a lot of nodes in one day, you might want to use more than one email.

any idea why when i did this the nodes were barley getting any storage saved on it?

You dont need VB to run multiple GUI nodes.
I hav 5 of them on 1 windows pc, you just need to know how to cook it.

How would i do this?

More nodes you run - less traffic will get.
All nodes behind the same /24 subnet of public IPs are treated as a one node for uploads and downloads, and as a separate ones for uptime and audits checks.
Each new node should be vetted. It will receive only 5% of potential traffic until got vetted. To be vetted on one satellite it should pass 100 audits from it. For the one node it usually should take at least a month.
In case of multisetup all that small amount of traffic is distributed between your nodes behind the same /24 subnet of public IPs, so the vetting will be in the same amount of time longer as a number of nodes.
So, better to run the next node when the previous one almost full, in this case the vetting process will not take forever.

Running more then 1 node at a time per computer gives no benefits, Less one node is full. It is better to run nodes in different locations if you run more then one node per IP address all the data will be shared between all of the nodes, and the vetting process will take even longer to complete.

This is really important and the first time I heard of this. I was thinking if I need to expand storage in the future I would add new nodes (perhaps by adding large NAS units as nodes). I think I have far more bandwidth available (1 Gbps) than storage (10 TB) and from the recent Town Hall it sounds like the rule of thumb is one hard drive per node.

Yes. But if you have a RAID already, you definitely can use it.
Just do not use it via network - the SQLite databases are not compatible with the network connected drives at all. Use only local connected drives.

1 Like

Why longer?

I look up in the code - looks like there are no dependencies on amount of data stored on node and traffic generated for audits any more. Auditor just pick up random nodes for check. It does not filter nodes by ip/subnet.
Also it is stated in Audit-V2 blueprint (which i think is already implemented): https://github.com/storj/storj/blob/91a480f5a01c0bdf6538da6d5cb860562c933476/docs/blueprints/audit-v2.md
Its designed to remove dependency on amount of data stored for audit frequency.

Sharing same ip (or ip subnet) reduce normal traffic to node of course due to ip filtering in data distribution module.
But it should not longer affect node audits as it was before.

I don’t remember seeing that in the change logs yet, so not sure it’s implemented. But even if it is, this still applies.

Using reservoir sampling means we have an equal chance to pick a segment for every node. Since every segment also audits 79 other nodes, we also audit other nodes. The chance of a node appearing in a segment pointer is proportional to the amount of data that the node actually stores. The more data that the node stores, the more chance it will be audited.

This basically means that 1 out of 80 audits is equally distributed across nodes, but the other 79 are still distributed based on the amount of data a node holds. It definitely helps in testing smaller nodes much more often, but it doesn’t make audits happen on all nodes with equal frequency.

Why oh why run multiple Windows 10 VM’s? The overhead must be huge. If you must use virtualization, create one Ubuntu VM and create multiple storagenodes with docker. Ubuntu and docker with 4 storagenodes runs rock solid here.

And don’t start all nodes at the same time, wait at least until a node is vetted before you start the next one. Otherwise the whole vetting process will take forever.