Namada custom

This repo emits a container image that can run in 4 modes. Running all 4 together in the setup described below results in a Namada node whose syncing can be paused. While syncing is paused, Undexer can reliably index the historical data that would normally be pruned from the chain after 2 epochs, and is thus only available during sync (see anoma/namada#3810).

Mode	Command	Networks	Requires init?	Description
node	control_node.js	⚠️Internal only	No	Manages the Namada node. MUST run in isolated network (no Internet access except through rpc-proxy and sync-proxy). Parses node log; when epoch increments, node messages sync-proxy to pause the sync.
rpc-proxy	rpc_proxy.js	Internal, External	⚠️Yes	Allows connections from indexer to node.
sync-proxy	sync_proxy.js	Internal, External	⚠️Yes	Allows connections from node to peers. By pausing the contained proxy, node sync is paused, so that the indexer can catch up.
node-status	status.js	Internal, External	No	Manages the other three. When indexing has caught up, Undexer tells node-status to resume sync, and node-status tells sync-proxy to restart the proxy.

Note that rpc-proxy and sync-proxy require an init process in the container (docker run --init, or init: true in docker-compose.yml) so that Docker is able to reap the zombie processes that are created when the internal simpleproxy is killed by the managing script.

Here's an example Docker Compose manifest which describes the relations between the containers. Adapt as needed.

volumes:
database:
networks:
database:
external:
internal:
internal: trueservices:
postgres:
image: postgres:16.2-alpinenetworks: [ database ]restart: unless-stoppedports: [ "127.0.0.1:5432:5432" ]environment:
- POSTGRES_PASSWORD=insecure
- POSTGRES_USER=postgres
- POSTGRES_DB=postgreshealthcheck:
test: ["CMD-SHELL", "pg_isready -U postgres -d postgres"]interval: 10stimeout: 5sretries: 5indexer:
entrypoint: /app/undexercommand: indexnetworks: [ external, database ]image: ghcr.io/hackbg/undexer:v4restart: unless-stoppeddepends_on: { postgres: { condition: service_healthy } }environment: { RPC_URL: "http://rpc-proxy:26657" }rpc-proxy:
entrypoint: /srv/rpc_proxy.jsnetworks: [ external, internal ]image: ghcr.io/hackbg/namada-for-undexer:maininit: truerestart: unless-stoppednode:
entrypoint: /srv/control_node.jsnetworks: [ internal ]image: ghcr.io/hackbg/namada-for-undexer:mainrestart: unless-stoppedsync-proxy:
entrypoint: /srv/sync_proxy.jsnetworks: [ external, internal ]image: ghcr.io/hackbg/namada-for-undexer:maininit: truerestart: unless-stoppedenvironment: { REMOTE: "namada-peer.mandragora.io:26656" }