Failed to run testnet node using docker image

[shanithkk]

Description

Couldn't run the tesnet node, failing due to "Failed to create RocksDB directory"

[Description of the issue or feature]

Error: Service(Client(Backend("Failed to create RocksDB directory: `Os { code: 13, kind: PermissionDenied, message: \"Permission denied\" }`.")))

Research/based on

Testnet

Reproduce
docker run centrifugeio/centrifuge-chain:latest-2b70c70-23-11-23 --chain=catalyst

[Which codebases, pallets or designs did you base your design on]

How will this affect the code base

[Does it improve readability, performance? Will it introduce new complexity?]

What are foreseen obstacles or hurdles to overcome?

[Are migrations needed, structural changes around testing etc?]

[mustermeiszer]

Hello @shanithkk are you an Axelar validator? Which instructions are you following?

[wischli]

Can you share your node startup command? While working on #1577, I became aware of the issue that the database type needs to explicitly set in the CLI (e.g. --database=auto) even though there exists a default value. This affects all commits starting at b9dffc0.

Moreover, the websocket CLI is not supported anymore (inherited from Polkadot v0.9.42). Please check our local docker compose file for a working command set:

centrifuge-chain/docker/docker-compose-local-chain.yml

Lines 18 to 29 in bcf8a05

	--chain="${PARA_CHAIN_SPEC:-development}"
	--alice
	--parachain-id="2000"
	--wasm-execution=compiled
	--execution=wasm
	--port=30333
	--rpc-port=9933
	--rpc-external
	--rpc-cors all
	--rpc-methods=Unsafe
	--log="main,info,xcm=trace,xcm-executor=trace"
	--database=auto

Please note that the image you are running is still experimental. The latest official one centrifugeio/centrifuge-chain:main-20231006174657-160318c7 is based on Polkadot v0.9.38 and does not requrie explicitly setting the db type.

Starting with the next official docker image, we will improve our tags to better distinguish between experimental and production clients. Moreover, the release notes (WIP) will contain exhaustive information about the breaking CLI changes.

[shanithkk]

Thank you its working with --database=auto, also may i know why bin/sh is removed. because i'm trying a new tool to spawn node its need a bin/sh. and also
Node logs showing something like
The bootnode you want to connect to at `/dns/polkadot-bootnode.polkadotters.com/tcp/30333/p2p/12D3KooWCgNAXvn3spYBeieVWeZ5V5jcMha5Qq1hLMtGTcFPk93Y` provided a different peer ID `12D3KooWPAVUgBaBk6n8SztLrMk8ESByncbAfRKUdxY1nygb9zG3` than the one you expect `12D3KooWCgNAXvn3spYBeieVWeZ5V5jcMha5Qq1hLMtGTcFPk93Y`.

[wischli]

Thank you its working with --database=auto, also may i know why bin/sh is removed. because i'm trying a new tool to spawn node its need a bin/sh.

The image you are running should have bin. We reverted the removal in #1551. It was removed beforehand to shrink the Docker image and reduce attack vectors.

and also Node logs showing something like The bootnode you want to connect to at `/dns/polkadot-bootnode.polkadotters.com/tcp/30333/p2p/12D3KooWCgNAXvn3spYBeieVWeZ5V5jcMha5Qq1hLMtGTcFPk93Y` provided a different peer ID `12D3KooWPAVUgBaBk6n8SztLrMk8ESByncbAfRKUdxY1nygb9zG3` than the one you expect `12D3KooWCgNAXvn3spYBeieVWeZ5V5jcMha5Qq1hLMtGTcFPk93Y`.

In the root comment, you said you want to connect to Catalyst which is a parachain on Rococo relaychain. However, the error here shows that you are pointing to Polkadot bootnodes.

Can you elaborate on what Centrifuge chain you want to run?

[shanithkk]

Thank you @wischli, actually we are building a new tool for polkadot ecosystem, that make the testing and deployment easier for relay chain and parachains. if it is possible can you make a default tag like latest for updating, so we don't worry about new tag anymore. now we actually changing the tag manually

[shanithkk]

Can you elaborate on what Centrifuge chain you want to run?

actually i was running the chain without mentioning the relay chain '-- --chain=rococo'. Thank you for pointing out @wischli