akash-network / docs Goto Github PK

The deploy guide does not

• have provider attributes that work on mainnet by default
• have anything that says you should look for bids and then create a lease

People have no idea what to do - it's totally broken documentation.

From Morgan:

I think that the beginning of the paper could use a rewrite in order to quickly explain to the reader why Akash is interesting and valuable. If the reader already believes that Akash is interesting and valuable and they want to understand it further, then this rewrite would not be useful. However, I think that it would be a good idea to explain and sell the idea of Akash at the beginning of every early Akash whitepaper because any of them could be somebody's introduction to Akash. When writing on a new topic it's usually a good idea not to assume the reader has any idea of what you are talking about. I would be happy to help with such a rewrite or even do it myself.

See here

What are the recommended specs to run an Akash node, provider, validator, etc.

Example:
2/4 cpus (cpus you can take as a normal aws/gcloud will use)
32/64 gb ram
1Tb + storage
high speed internet connection

I would love to see an high level architecture example picture how all parts (API, RPC, Validator, Node, protocols, ... work together)

In sdl, disk has been replaced with storage.

Add the akash provider lease-events command; it's similar to lease-logs.

Both of these should be in a section about "debugging", or, at the very least, they need to be before the deployment is closed.

getting logs is done through akash provider lease-logs. the documentation has service-logs.

The following link is no longer active:

Before setting up your validator node, make sure you've already gone through the Full Node Setup guide

https://github.com/ovrclk/docs/blob/v0.8/validator-guides/create-a-testnet-validator.md

we used to use u to denote micro akash; that unit is no longer valid. it should be akash always and everywhere.

The command to create your client key seems to be missing from the deployment guide.
akash tx cert create client --from $KEY_NAME --chain-id $AKASH_CHAIN_ID

Without it the akash tx deployment create command fails with "Error: no certificate file found for account"

This is using
AKASH_VERSION=v0.10.0
AKASH_CHAIN_ID=akashnet-2-testnet-2

Install guide link and testnet guide link are not working in client doc.

Akash is, first and foremost, a platform that allows clients to procure resources from providers. This is enabled by a blockchain-powered distributed exchange where clients post their desired resources for providers to bid on. The currency of this marketplace is a digital token, the AKT, whose ledger is stored on a blockchain.

[peter@nixos:~]$ curl -I https://akash.network/l/econ-paper
HTTP/2 404
date: Mon, 31 Aug 2020 18:27:03 GMT
content-type: text/html; charset=UTF-8
set-cookie: __cfduid=da52c50d8138781775a99fa90422a3c6c1598898423; expires=Wed, 30-Sep-20 18:27:03 GMT; path=/; domain=.akash.network; HttpOnly; SameSite=Lax; Secure
expires: Wed, 11 Jan 1984 05:00:00 GMT
cache-control: no-cache, must-revalidate, max-age=0
link: <https://akash.network/wp-json/>; rel="https://api.w.org/"
via: 1.1 vegur
cf-cache-status: DYNAMIC
cf-request-id: 04e75edd620000c9a1f50ca200000001
expect-ct: max-age=604800, report-uri="https://report-uri.cloudflare.com/cdn-cgi/beacon/expect-ct"
server: cloudflare
cf-ray: 5cb8cda89c3ec9a1-SEA

So that people know that they can report issues, and they report it to the right place.

Add an FAQ section

• Does it work with GPUs?
• How much money will I make?
• When supermini?
• etc...

https://github.com/ovrclk/engineering/issues/33

when i do init where this weird chain id coming from in query output ?
"chain_id": "test-chain-uU6Is4",
"gentxs_dir": "",

• Example Repo

If someone is setting up a provider they shouldn't need to handle real AKT on mainnet just to test if their provider works.

Should add a no-frills guide to running the example from _run/kube etc. to run a local node & provider to verify the kubernetes cluster they want to sell on the network is ready for mainnet.

pruning="everything"

should not be used. let's warn people about it. it causes problems like this: https://github.com/ovrclk/akash/issues/1206

We need to add docs & example for new provider lease-shell command. It should show how to use the command on an example deployment, both for interactive shell and for other actions (i.e. copying a file from a docker image)

We now have three home directories in the wild: ~/.akashd and ~/.akashctl from 0.7.x and ~/.akash from 0.9.x.

Write an upgrade how-to, especially for:

• migrating keyring
• replacement for client config
• migrating node config, data

Git checkout version command is missing in install doc

The below sections needs an update:

• Query Network
• Send Tokens
• Get remote node status

In documentation here https://docs.akash.network/testnet-challenges/testnet-challenges/guided-deployments#challenge-2-week-2

SDL link leading in to SDL documentation: https://docs.akash.network/documentation/sdl

Expected behavior:

Lead do deploy-2-2.yaml
https://github.com/ovrclk/docs/blob/ac011af1802a7c5fc7bb90d979c4f4877eaa24e1/testnet-challenges/deploy-2-2.yaml

I've been learning about Akash network. I have a couple of questions,

I hope these details can be documented,

what SLSA guarantees does providers offer? how does Akash confirm the clients, the providers are indeed running what they are expected to run?

Is it possible for multiple wallets to own deployments? a multisig mode so that deployments can be decentralized in terms of ownership. is it possible to have DAO governance to update the signature set?

please direct me to appropriate channels to get them clarified.

We've added to akash the ability to do custom bid pricing with an external script. The command line switch document how to set this up, but do not actually document the exact interface the script needs to implement.

We need to

• Document the interface
• Document any data structures used
• Document any expectations around the script
• Provide a complete example script in a common scripting language e.g. python3

Check back on June 17th, 2021.

it's June 30th now 🤷

We've got two provider guides

https://github.com/ovrclk/docs/blob/master/guides/provider.md

https://github.com/ovrclk/docs/blob/master/guides/provider/README.md

We probably need to try and deduplicate this. also the steps to create the provider server cert is missing from at least one guide.

I recommend setting up a HackerOne bug bounty to officially and properly reward hackers who find security vulnerabilities and bugs.

Currently, we have a limited "bug bounty" in our documentation at https://docs.akash.network/support:

Security Issues & Bug Bounties
Akash will pay for finding security issues. If you find a security issue or bug and wish to report it to the core team privately, please send an email with all of the details.
Email [email protected]

How about add hardware requirements of node/validator to doc? Users often ask this.

https://github.com/ovrclk/docs/blob/master/operator/node/README.md

People are not creating their certificates.

We should strongly encourage people to enable snapshotting and state sync. somewhere around pruning of state and fast sync.

1. have the instructions enable snapshotting
2. we should add a new file to the network configuration at ovrclk/net state-sync-peers.txt and tell people to skip this step until that is populated.

dehazelabs.com are stuck on bidding as a Provider on Testnet. They aren't able to find instructions on how to do this. They have been stuck for a week.

The CEO of https://dehazelabs.com reached out on Discord and wants to be a provider and a tenant. This is the right kind of Provider we want to onboard early. Their timeline to go on Mainnet is ~4 weeks.

“yeah we are not able to find any doc to be able to bid as a provider on the testnet. Yeah we followed this and there were a lot of missing instructions. We have notes which I will share so you guys can update the docs. "

When running "akash tx staking create-validator {...}" with --gas="auto" an out of gas error is returned. Running the command without --gas="auto" succeeds.

-side note: the docs are not clear on how to get $VALIDATOR_ADDRESS for the section "View Validator Description" per discord discussion it can be found in the return message of a successful create-validator command.

As an akash user I should be able to join the testnet as a validator

As an akash user I should be able to deploy an app on the testnet

args is pretty unclear and surprising. With the following SDL:

  web:
    image: raviforbole/video-chat
    args:
      - "--VUE_APP_SOCKET_HOST=http://localhost"
      - "--VUE_APP_SOCKET_PORT=3000"

Pod creation fails with:

f7shd5o2mbfnpgmlij0af111naf7nif46al3l3gsvob7k   20s         Warning   Failed              pod/web-6bff748d8-lrgs9        Error: failed to start container "web": Error response from daemon: OCI runtime create failed: container_linux.go:349: starting container process caused "exec: \"--VUE_APP_SOCKET_HOST=http://localhost\": stat --VUE_APP_SOCKET_HOST=http://localhost: no such file or directory": unknown

So it looks like the first element of args: should be the path to the executable.

example:

 curl https://raw.githubusercontent.com/ovrclk/akash/master/godownloader.sh | sh -s -- v0.6.1

There is additionally a bug in the booster script at https://github.com/arno01/akash-tools/blob/a9b7b09f167a65ee269185c5599825aa04b9143b/cli-booster/akash.source#L656-L657 where both instances of -ge should be -gt or the second variable clobbers the first comparing equal with 0.

Add or update usage documentation for CLI Top-level commands and subcommands

• config
• help
• keys
• query
• rest-server
• status
• tx
• version

Requirements

1. Each top-level command should be in a separate file with the name <command>.md
2. Each command should have the below structure. See keys and keys add for reference

• Description
• Usage
• Example
• Arguments
• Flags