Root cause: When mpc-service restarts, the old Kafka consumer doesn't
properly disconnect, causing the broker to wait for sessionTimeout
(~5 minutes) before completing rebalance. This blocks app startup.
Solution: Enable NestJS shutdown hooks with app.enableShutdownHooks().
This ensures onModuleDestroy() is called on SIGTERM/SIGINT, which calls
consumer.disconnect() and allows immediate rebalance on next startup.
Also reverted the "don't await consumer.run()" workaround since the
proper fix is graceful shutdown.
Sources:
- https://github.com/tulios/kafkajs/issues/807
- https://kafka.js.org/docs/consuming🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
consumer.run() never resolves as it runs continuously. Awaiting it
blocks onApplicationBootstrap which prevents app.listen() from being
called, causing the service to never start listening on port 3006.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit fixes three critical bugs that prevented the wallet creation
flow from completing successfully:
1. mpc-service: extraPayload not included in Kafka messages
- KeygenCompletedEvent's extraPayload (containing userId, accountSequence,
username, derivedAddresses) was being set dynamically but not serialized
- identity-service received events without userId and skipped processing
- Fix: Merge extraPayload into the published payload in event-publisher
2. mpc-service: KAFKA_BROKERS hostname mismatch
- mpc-service used KAFKA_BROKERS=rwa-kafka:29092
- Kafka advertises itself as kafka:29092 in cluster metadata
- During consumer group rebalance, mpc-service couldn't connect to
the coordinator address returned by Kafka
- Fix: Use kafka:29092 to match Kafka's advertised listener
3. blockchain-service: recovery_mnemonics table missing
- RecoveryMnemonic model exists in schema.prisma but not in migration
- prisma migrate deploy found no pending migrations
- Address derivation failed with "table does not exist" error
- Fix: Use prisma db push instead of migrate deploy to sync schema
Tested: E2E flow now completes successfully
- POST /user/auto-create creates account
- MPC keygen completes and publishes event with extraPayload
- blockchain-service derives addresses and saves recovery mnemonic
- GET /user/wallet returns status=ready with 3 addresses and mnemonic
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Changes across all three services to properly associate recovery mnemonics
with account sequence numbers instead of user IDs, following DDD principles:
identity-service:
- Add accountSequence to MpcKeygenRequestedEvent payload
- Pass accountSequence when publishing keygen request
- Remove direct access to recoveryMnemonic table (now in blockchain-service)
- Call blockchain-service for mnemonic backup marking
- BlockchainWalletHandler no longer saves mnemonic (stored in blockchain-service)
mpc-service:
- Add accountSequence to KeygenRequestedPayload interface
- Pass accountSequence through to blockchain-service when deriving addresses
- Include accountSequence in KeygenCompleted event extraPayload
blockchain-service:
- Add accountSequence to derive-address API and internal interfaces
- Add accountSequence to KeygenCompletedPayload extraPayload
- Add PUT /internal/mnemonic/backup API for marking mnemonic as backed up
- Store recovery mnemonic with accountSequence association
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Increase sessionTimeout from 30s to 5 minutes
- Increase heartbeatInterval from 3s to 10s
- Add rebalanceTimeout of 5 minutes
- This prevents consumer from being kicked out during long MPC keygen operations
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
blockchain-service uses global API prefix, need to include it in URLs.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
mpc-account-service expects JWT Bearer tokens, not X-API-Key header.
Added JWT token generation and use MPC_JWT_SECRET env var.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Changed from MPC_SYSTEM_URL to MPC_ACCOUNT_SERVICE_URL to match docker-compose config.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add BlockchainClientService to call blockchain-service /internal/derive-address
- Call derive addresses after keygen completes with MPC public key
- Include derived addresses (BSC, KAVA, DST) in keygen completed event
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add comprehensive [INIT], [CONNECT], [PUBLISH], [RECEIVE], [HANDLE] logs
to identity-service and mpc-service Kafka services
- Add KeygenStarted event for tracking keygen progress
- Add MpcKeygenCompletedHandler to process keygen completion events
- Fix topic routing for MPC events between services
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Replace synchronous HTTP polling with Kafka event-driven model for MPC operations:
- Add MPC event consumer service in mpc-service for keygen/signing requests
- Add keygen-requested and signing-requested event handlers
- Add MPC event consumer in identity-service for completion events
- Extend mpc-client.service with async event-driven methods
- Support backward compatibility via MPC_USE_EVENT_DRIVEN env var
Topics: mpc.KeygenRequested, mpc.SigningRequested, mpc.KeygenCompleted,
mpc.SigningCompleted, mpc.SessionFailed
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add MPCController with keygen and sign endpoints
- Add MPCCoordinatorService to coordinate with mpc-system
- Add MpcWallet, MpcShare, MpcSession entities for data storage
- Update identity-service mpc-client to call mpc-service
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Remove Address field from PartyEndpoint (parties connect to router themselves)
- Update K8s Discovery to only manage PartyID and Role labels
- Add Party registration and SessionEvent protobuf definitions
- Implement PartyRegistry and SessionEventBroadcaster domain logic
- Add RegisterParty and SubscribeSessionEvents gRPC handlers
- Prepare infrastructure for party-driven MPC coordination
This is the first phase of migrating from coordinator-driven to party-driven
architecture following international MPC system design patterns.
将 `import WebSocket from 'ws'` 改为 `import * as WebSocket from 'ws'`
以兼容 CommonJS 模块系统
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
hailin
Developer