Update DAL file server tutorial for Ghostnet

[timothymcmackin]

Update the DAL file server tutorial:

• Use Ghostnet instead of Weeklynet for convenience
• Run a layer 1 node as part of the environment to get around an ACL issue on some public RPC nodes
• Little clarifications and corrections
• Reduce reliance on setting environment variables because we're constantly opening new windows where they might not be set
• Switch to --observer-profiles to be in observer mode for the DAL node
• More info about what you see happening on Explorus
• Add info about what happens if there are no active attesters
• Add --legacy more for the DAL trusted setup
• Fix Rust version to 1.80 because the kernel doesn't compile at version 1.81.

Preview: https://docs-staging-git-update-dal-file-tutorial-trili-tech.vercel.app/tutorials/build-files-archive-with-dal

As noted in the comments, this tutorial doesn't currently work on Ghostnet because of an issue with DAL attesters. However, we may want to merge this update now anyway because it no longer works consistently on Weeklynet.

There is a note that says:

For the first 4 Tezos blocks produced after the origination of the Smart Rollup, the kernel will report that no slot has been attested for the targeted level, even if Explorus states the opposite. This is because, as of January, 2024, a Smart Rollup cannot fetch the content of a slot published before it is originated. This is why you must wait for 4 blocks before seeing slot page contents being logged.

Should this be 8 blocks for Ghostnet because that's the attestation_delay?

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated (UTC)
docs-staging	✅ Ready (Inspect)	Visit Preview	Jan 29, 2025 1:16pm

[ezal]

The changes look ok to me.

I haven't been able to finish the tutorial, as expected:

• latest doesn't yet work with the current bootstrap node on Ghostnet
• slots are not attested because there are not enough DAL-enabled bakers on Ghostnet.

But I guess this PR is a preparation for the case when these two issues will be fixed.

[timothymcmackin]

I followed this tutorial on 27 January and it seems to work on Ghostnet. Can anyone else review? I think this update may finally be ready to go. @ezal @francoisthire

[francoisthire]

I think the octez-client command above also requires this assumption. I suggest you move it before the first command that requires it.

[timothymcmackin]

Added this:

The Octez client assumes that your local node is running at http://127.0.0.1:8732. If your node is running at a different host name or port, pass the host name and port of the node to the --endpoint argument. For example, if the node is running on port 8733, include --endpoint http://127.0.0.1:8733 in the command.

[francoisthire]

Nit: I don't understand why sometime the endpoint is specified sometimes it is not. My guess would be than when it is not needed as for octez-client, you don't specify it, but for the octez-smart-rollup-node there might be no default so it needs to be specified?

Why not putting the endpoint everywhere so that a reader can copy/paste it easily and modify it if it uses a different endpoint?

[timothymcmackin]

It's klunky and unnecessary to put the endpoint in every command. I'll remove the rest of them unless you really think people should specify --endpoint every time.

[saroupille]

I am fine with it.