NagyZoltanPeter e8566db8ee
fix(persistency): own Persistency per node instead of a process-global singleton (#4109)
* fix(persistency): own Persistency per node instead of a process-global singleton

The Persistency singleton (gPersistency) was allocated on whichever FFI
thread first ran waku.start; under --mm:refc its memory belonged to that
thread's heap, so a second library context adopting it read a foreign
heap (SIGSEGV in sdsPersistence -> openJob -> tables.rawGet), one
context's stop stole the other's SDS persistence, and a destroyed
context poisoned re-init with a different local-storage-path.

- remove the singleton (instance/reset/gPersistency); Persistency.new is
  the public constructor, no lock needed (instances are thread-confined)
- own the instance as Waku.persistency: created and provided in
  waku.start, cleared and closed in waku.stop on the owning thread
- expose it via a sync GetPersistency RequestBroker scoped to the node's
  BrokerContext; sdsPersistence resolves through it (same-thread ref
  return, no marshalling)
- add InMemoryStoragePath (":memory:") support: private in-memory SQLite
  per job worker, for tests
- rewrite test_singleton as per-instance + broker coverage; rewrite
  test_thread_affinity from a known-failing UB repro into a regression
  guard (two in-memory jobs, worker spinup/teardown, cross-thread broker
  denial) and register it in test_all; migrate remaining tests to
  new/close; the FFI lifecycle test's stop-steals and different-paths
  cases now pass against the real dylib

The FFI destroy-without-stop teardown gap remains tracked in #4108.

Fixes #4103

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(persistency): create the instance only after startup fully succeeds

waku.start has many error return paths; creating Persistency early meant
every one of them left the field set and the GetPersistency provider
installed with no teardown. Persistency.new is inert (no threads or
files until the first openJob) and every consumer runs post-start, so
creating and providing it as the last startup step removes the need for
any error-path cleanup entirely.

Addresses PR #4109 review feedback.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(persistency): create early for startup-stage restores, tear down on failed start

Creating the instance as the last startup step made it impossible for
any stage of start to restore persisted data (e.g. a future store-state
restore). Restore the original ordering -- create and provide the
instance first -- and cover every error return path of waku.start with a
success-flag defer that clears the provider and closes the instance.
Teardown is factored into closePersistency, shared by stop and the
failed-start path.

Addresses PR #4109 review discussion.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* chore(persistency): address remaining Copilot review findings

- document that Persistency instances are thread-confined (not
  thread-safe) on the type itself, pointing at the GetPersistency broker
  as the sanctioned access path
- use tryRemoveFile for the FFI test's log cleanup so an unremovable
  file cannot fail the test for unrelated reasons

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* Fix comment

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-08-06 23:35:10 +02:00
..

Waku

This folder contains code related to Waku, both as a node and as a protocol.

Introduction

This is an implementation in Nim of the Waku suite of protocols.

See specifications.

How to Build & Run

Prerequisites

  • GNU Make, Bash and the usual POSIX utilities. Git 2.9.4 or newer.

Wakunode binary

# The first `make` invocation will update all Git submodules.
# You'll run `make update` after each `git pull`, in the future, to keep those submodules up to date.
make wakunode2

# See available command line options
./build/wakunode2 --help

# Connect the client directly with the Status test fleet
# TODO NYI
#./build/wakunode2 --log-level:debug --discovery:off --fleet:test --log-metrics

Note: building wakunode2 requires 2GB of RAM. The build will fail on systems not fulfilling this requirement.

Setting up a wakunode2 on the smallest digital ocean droplet, you can either

  • compile on a stronger droplet featuring the same CPU architecture and downgrade after compiling, or
  • activate swap on the smallest droplet, or
  • use Docker.

Waku Protocol Test Suite

# Run all the Waku tests
make test

To run a specific test file or test case:

# Run all tests in a specific file
make test tests/waku_filter_v2/test_waku_filter.nim

# Run a specific test case within a file
make test tests/waku_filter_v2/test_waku_filter.nim "specific test name"

Alternatively, you can invoke the Nim compiler directly. For more on available flags, refer to the compiler flags and chronicles documentation.

nim c -r -d:chronicles_log_level=WARN --verbosity=0 --hints=off ./tests/waku_filter_v2/test_waku_filter.nim

You may also want to change the outdir to a folder ignored by git.

nim c -r -d:chronicles_log_level=WARN --verbosity=0 --hints=off --outdir=build ./tests/waku_filter_v2/test_waku_filter.nim

Waku Protocol Example

There are basic examples of both publishing and subscribing, more limited in features and configuration than the wakunode2 binary, located in examples/.

There is also a more full featured example in apps/chat2/.

Using Metrics

Metrics are available for Waku nodes.

make wakunode2
./build/wakunode2 --metrics-server

Ensure your Prometheus config prometheus.yml contains the targets you care about, e.g.:

scrape_configs:
  - job_name: "waku"
    static_configs:
      - targets: ['localhost:8008', 'localhost:8009', 'localhost:8010']

For visualisation, similar steps can be used as is written down for Nimbus here.

There is a similar example dashboard that includes visualisation of the envelopes available at metrics/waku-grafana-dashboard.json.

Spec support

All Waku RFCs reside at rfc.vac.dev. Note that Waku specs are titled WAKU2-XXX to differentiate them from a previous legacy version of Waku with RFC titles in the format WAKU-XXX. The legacy Waku protocols are stable, but not under active development.

Generating and configuring a private key

By default a node will generate a new, random key pair each time it boots, resulting in a different public libp2p multiaddrs after each restart.

To maintain consistent addressing across restarts, it is possible to configure the node with a previously generated private key using the --nodekey option.

wakunode2 --nodekey=<64_char_hex>

This option takes a Secp256k1 private key in 64 char hexstring format.

To generate such a key on Linux systems, use the openssl rand command to generate a pseudo-random 32 byte hexstring.

openssl rand -hex 32

Example output:

$ openssl rand -hex 32
6a29e767c96a2a380bb66b9a6ffcd6eb54049e14d796a1d866307b8beb7aee58

where the key 6a29e767c96a2a380bb66b9a6ffcd6eb54049e14d796a1d866307b8beb7aee58 can be used as nodekey.

To create a reusable keyfile on Linux using openssl, use the ecparam command coupled with some standard utilities whenever you want to extract the 32 byte private key in hex format.

# Generate keyfile
openssl ecparam -genkey -name secp256k1 -out my_private_key.pem
# Extract 32 byte private key
openssl ec -in my_private_key.pem -outform DER | tail -c +8 | head -c 32| xxd -p -c 32

Example output:

read EC key
writing EC key
0c687bb8a7984c770b566eae08520c67f53d302f24b8d4e5e47cc479a1e1ce23

where the key 0c687bb8a7984c770b566eae08520c67f53d302f24b8d4e5e47cc479a1e1ce23 can be used as nodekey.

wakunode2 --nodekey=0c687bb8a7984c770b566eae08520c67f53d302f24b8d4e5e47cc479a1e1ce23

Configuring a domain name

It is possible to configure an IPv4 DNS domain name that resolves to the node's public IPv4 address.

wakunode2 --dns4-domain-name=mynode.example.com

This allows for the node's publicly announced multiaddrs to use the /dns4 scheme. In addition, nodes with domain name and secure websocket configured, will generate a discoverable ENR containing the /wss multiaddr with /dns4 domain name. This is necessary to verify domain certificates when connecting to this node over secure websocket.

Using DNS discovery to connect to existing nodes

A node can discover other nodes to connect to using DNS-based discovery. The following command line options are available:

--dns-discovery              Enable DNS Discovery
--dns-discovery-url          URL for DNS node list in format 'enrtree://<key>@<fqdn>'
--dns-addrs-name-server  DNS name server IPs to query. Argument may be repeated.
  • --dns-discovery is used to enable DNS discovery on the node. Waku DNS discovery is disabled by default.
  • --dns-discovery-url is mandatory if DNS discovery is enabled. It contains the URL for the node list. The URL must be in the format enrtree://<key>@<fqdn> where <fqdn> is the fully qualified domain name and <key> is the base32 encoding of the compressed 32-byte public key that signed the list at that location.

A node will attempt connection to all discovered nodes.

This can be used, for example, to connect to one of the existing fleets. Current URLs for the published fleet lists:

  • production fleet: enrtree://AIRVQ5DDA4FFWLRBCHJWUWOO6X6S4ZTZ5B667LQ6AJU6PEYDLRD5O@sandbox.waku.nodes.status.im
  • test fleet: enrtree://AOGYWMBYOUIMOENHXCHILPKY3ZRFEULMFI4DOM442QSZ73TT2A7VI@test.waku.nodes.status.im

See the separate tutorial for a complete guide to DNS discovery.

Enabling Websocket

Websocket is currently the only Waku transport supported by browser nodes that uses js-waku. Setting up websocket enables your node to directly serve browser peers.

A valid certificate is necessary to serve browser nodes, you can use letsencrypt:

sudo letsencrypt -d <your.domain.name>

You will need the privkey.pem and fullchain.pem files.

To enable secure websocket, pass the generated files to wakunode2: Note, the default port for websocket is 8000.

wakunode2 --websocket-secure-support=true --websocket-secure-key-path="<letsencrypt cert dir>/privkey.pem" --websocket-secure-cert-path="<letsencrypt cert dir>/fullchain.pem"

Self-signed certificates

Self-signed certificates are not recommended for production setups because:

  • Browsers do not accept self-signed certificates
  • Browsers do not display an error when rejecting a certificate for websocket.

However, they can be used for local testing purposes:

mkdir -p ./ssl_dir/
openssl req -x509 -newkey rsa:4096 -keyout ./ssl_dir/key.pem -out ./ssl_dir/cert.pem -sha256 -nodes
wakunode2 --websocket-secure-support=true --websocket-secure-key-path="./ssl_dir/key.pem" --websocket-secure-cert-path="./ssl_dir/cert.pem"

Enabling QUIC

QUIC is a UDP-based transport that peers can use to connect to your node.

The default port for QUIC is 60000.

wakunode2 --quic-support=true

To listen on a different UDP port, use --quic-port:

wakunode2 --quic-support=true --quic-port=<port>