We’ve been poking at various bits of the network recently, and have something that feels reasonably stable, but also maaayyyyy still have some issues. And we’d love your help in checking this out!
We’re not calling this a testnet as yet… the CLI install process is not as tight as we’d like. So until we have that in place, and can offer support for that, this playground is for intrepid community testnet type folk. Also, with it being a Friday afternoon/evening and so close to Christmas, team availability is low so we can only offer limited support.
So please don’t expect a wildly stable testnet, that will come in due course. We still need to get into fixing of membership and deep dive in DKG, although these seem stable enough for small networks, we are just not convinced they are fully correct yet.
But to help us we would like to try this out and surface bugs still lurking about. It hopefully also gives those itching for a testnet something to play with.
As things stand, DKG is looking much healthier, but we’re working on integrated BRB membership just now. As such, while you can join with a node (yay decentralisation!), we’re not aiming to look at this just now. And so won’t be providing any guide for that here.
What we want to focus on here is testing the new and updated CLI release, i.e the safe files put and safe files cat commands. (With the env var RUST_LOG=safe_network=trace, if you want to get some logs printed out).
Can we put data, can we retrieve it?
We’ve seen a few bugs that we’re not sure if they occur solely on the continuous integration machines as a result of power limitations there. (Those machines have 2 vCPUs and 7gb ram… and on which we boot up many many nodes). So we want to know if we can surface these in the wild.
We’re looking to see if we still see NoAdult errors being returned to clients, or indeed if we see any Database(NotEnoughSpace) errors at nodes (we’re hoping not to see either of those).
To help in this goal, the nodes forming the playground are of limited capacity (1gb), and we expect to blow through that resonably quickly (if @southside is involved at least 
). If everything goes well, we may look to put up a larger capacity playground to keep poking at the new CLI…
In the meantime we ask that anyone participating keeps the sizes of files they upload relatively small, perhaps a max of a 100mb, to allow the network to stay active and usable for long enough for us to gather adequate information to inform us on the next steps.
On the CLI side, we’re hoping some new fixes including a reimplemented chunk cache, and adjusted self-verifying PUT should increase the stability of uploads for us now. (One idea we’re testing is that the CLI didn’t always wait for chunks to go through the full AE flow.)
This should impact larger file uploads more. But will also increase the time taken for that upload via CLI, as we now verify every chunk is PUT to the network before we terminate the upload.
We’ve experienced some longer times for uploading and/or downloading Blobs from our testnets for some of us, therefore, if you face this same issue with your connection, you can increase the CLI’s timeout (default is 2 mins), e.g. to set it to 4 mins use:
$ export SN_CLI_QUERY_TIMEOUT=240
So with all of the above caveats, anyone who does want to play with a live network can now:
- remove
~/.safe - run
curlhttps://raw.githubusercontent.com/maidsafe/safe_network/master/resources/scripts/install.sh| bashto install the CLI - run
safe --versionto make sure you’re on0.39.1 - run
safe networks add playground https://safe-testnet-tool.s3.eu-west-2.amazonaws.com/public-node_connection_info.config safe networks switch playground
And you should be good to go 
P.S.
Instructions for the files put , cat and get commands can be found below:
ducklings
rice field
