[1.8.0] Release TODOs and cross-repository integration checklist

[imbajin]

Release scope

Use the current mainline of each repository to check HugeGraph 1.8.0 functionality, compatibility, documentation, and cross-repository integration: master for Server, Toolchain, and Doc; main for AI

The 1.7 → 1.8 path applies only to upgrade compatibility testing

Note

Key areas:

• HugeGraph Server: deployment, ToplingDB, Java/Groovy/TinkerPop upgrades, HStore ordered scans, and stability
• HugeGraph Toolchain: Hubble V2, Client, Loader, and SeaTunnel integration
• HugeGraph AI: GraphMCP / Skills, Python SDK, Text2Gremlin / GraphRAG
• Website and documentation: 1.8 code-documentation synchronization, release notes, and downloads
• Cross-repository integration and multi-architecture artifacts

1. HugeGraph Server

• Update the repository-wide version to 1.8.0 and synchronize the related documentation

Kubernetes / Helm / Docker

• Complete Helm/Kubernetes support
  • Complete Server #3218 and the corresponding Doc #494
  • Validate installation, upgrades, persistence, and failure recovery for the components covered by the chart, and clarify coverage of Hubble and other components
• Validate standalone, minimal HStore, and the HA deployment configurations supported by this release
  • Cover cold starts, health checks, persistence, restarts, PD leader changes, Store replacement, and DNS/connection recovery
  • Validate rebuilding a Store with an empty volume at its previous address, retiring the old Store ID, and manual recovery after fixing failures; document cases requiring operator intervention
  • Verify Server user authentication, PD service authentication, and Hubble Operations credentials separately

HStore / query correctness and stability

• Validate correctness and performance of HStore range-index ordered scans
  • Cover multiple partitions, filtering, limit/offset/page, cursor recovery, and duplicate/missing results; document the supported scope
  • Verify that the new protocol is enabled only after all Store nodes are upgraded, including rejection behavior with mixed versions
• Run core query and distributed stability regressions
  • Cover range/nested predicates, count, Cypher API authentication and failure responses, cross-Server Schema Cache synchronization, writes, and snapshot recovery

ToplingDB support

• Confirm release inclusion and validate ToplingDB integration: PR #3134 and integration and testing issue summary #240
  • Resolve ToplingDB integration issues and investigate unclassified findings tracked in task index-label id has conflicts with primitive index label #240; track general issues separately and distinguish merge readiness from release readiness
  • Verify explicit opt-in configuration, native library loading, and compatibility with the default RocksDB path across Server, PD, and Store
  • Check existing-data compatibility, dependencies, and LICENSE/NOTICE; document the current Linux x86_64 support boundary and deployment requirements

Java 17 / Groovy 4 / TinkerPop upgrades

• Advance the runtime and query-engine upgrade: PR #3117 and tracking issue #3069
  • Confirm the TinkerPop target for 1.8: the current PR baseline is 3.7.6, while 3.8 is follow-up work in the tracking issue; align implementation and acceptance scope if 3.8 is selected for this release
  • Complete Java 17 and Groovy 4 adaptations for builds, CI, images, and startup, and validate the script sandbox
  • Run Gremlin semantics, serialization, and Client compatibility regressions, cover PD/Store/HStore operation, and update upgrade documentation

Upgrade compatibility

• Exercise 1.7 → 1.8 data/configuration upgrades, backups, and rollback
  • Verify existing data and Schema, init_store.enabled, authentication configuration, and retained keys; check metadata cluster configuration to avoid apparently empty Schema after an upgrade
  • Update migration guidance for removed backends and the support matrix for RocksDB, HStore, and HBase

2. HugeGraph Toolchain

• Update the repository-wide version to 1.8.0 and synchronize the related documentation

Hubble

• Validate the Hubble V2 multilingual and permission workflows
  • Cover standalone/PD modes, authentication enabled/disabled, GraphSpace, default roles, unauthorized-access rejection, and consistent REST/Gremlin identities
  • Verify migration of existing H2 data, import-task recovery after restart, query results, and English/Chinese workflows

Client / Loader

• Run Java Client and Loader compatibility and import regressions against the Server RC
  • Cover FILE/HDFS/JDBC/Kafka, Graph Source, Schema copying, and concurrent date handling
  • Investigate the parser-threads > 1 NPE findings and either fix them or document verified support limitations; track #710
  • Rerun Hubble/Loader after Client changes and retain tests for older versions explicitly supported by the release

Apache SeaTunnel 3.0 support

• Complete HugeGraph Source/Sink integration: connector source
• Synchronize the integration documentation in the Doc repository: English guide and Chinese guide

HugeGraph-side integration and documentation are complete, with no additional work pending here; only the official SeaTunnel 3.0 release is awaited

This is an external release dependency, not an unfinished HugeGraph integration or acceptance task

3. HugeGraph AI

• Update the repository-wide version to 1.8.0 and synchronize the related documentation

GraphMCP / Skills / Python SDK

• Merge GraphMCP and the accompanying Skills into main: merge commit
• Validate release installation and upgrades for the Python SDK, GraphMCP, and AI applications
  • Prepare compatible artifacts in hugegraph-python → hugegraph-mcp order and verify wheel/sdist installation, MCP startup, and Skills in a clean environment, following the MCP release guide
  • Synchronize release versions, dependency constraints, lockfiles, images, and documentation; verify actual package availability rather than relying on source version declarations
• Connect Python Client, LLM, and GraphMCP to the same Server RC and validate authentication and GraphSpace integration
  • Cover version detection, non-default GraphSpace, Schema/vertices/edges/queries, and unauthorized-access rejection
• Complete 1.8 RC acceptance for GraphMCP and its Skills
  • Verify tool discovery, read-only defaults, structured queries, and expiration, replay prevention, and restart recovery in dry-run → plan_id → confirm → status/reconcile
  • Document supported confirmed writes, disabled operations, and the single SQLite writer-instance boundary; validate against actual Server RC capabilities using the MCP capability and safety boundaries

Text2Gremlin / GraphRAG

• Run core GraphRAG, Text2Gremlin, and graph extraction regressions
  • Cover PDF input/chunking, empty examples, invalid input, ID consistency, Schema persistence, no duplicate execution during streaming, and MCP/API credential isolation and response contracts
  • Record contract-test results separately from smoke tests against real model services
• Define the release scope and artifacts for Text2Gremlin data generation and fine-tuning on the text2gql branch
  • This capability is on text2gql; do not describe it as already available on mainline

4. Website and documentation

• Update the repository-wide version to 1.8.0 and synchronize the related documentation

Central tracker: Doc #491: code-doc sync before 1.8.0, pinned in the Doc repository

• Complete the Server, AI, and Computer/Vermeer documentation items already closed in monitor: save counters of each schema to variables #491: merge commit
• Complete the remaining synchronization work in #491
  • Audit PRs intended for 1.8 and record documentation impact, owners, and English/Chinese update links, or an explicit no-documentation-needed decision
  • Align Loader Kafka skipped_line documentation with runtime behavior, define support/removal/migration, and update the English and Chinese guides
• Correct PD authentication documentation and execute the English/Chinese deployment instructions against the RC
  • Remove outdated claims in the PD guide that passwords are unchecked or arbitrary values work; verify service identities, secrets, and authentication exemptions
  • Pin matching versions in Compose/Hubble examples and synchronize backend migration, HStore metadata prefixes and recovery boundaries, and Hubble V2 instructions
  • Add English/Chinese entry points for GraphMCP and Skills installation, permissions, and controlled writes
• Complete English/Chinese 1.8.0 release notes, download metadata and signature links, version navigation, and release-validation configuration
• Run existing website checks for builds, links, artifacts, version switching, and English/Chinese navigation
  • Validate the release snapshot using the updated OINK toolchain and existing strict/browser/version checks

5. Other TODOs

Cross-repository integration and multiple architectures

• Define release scope, owners, and compatibility by repository, and freeze commits/tags and RC artifacts
  • Align Server, Toolchain, and AI release versions and dependencies, with matching Doc version information; manage internal artifacts through each repository's release workflow
• Use one consistent RC set for cross-repository testing, replacing old services, floating master builds, and latest images
  • Prepare artifacts in dependency order: Server shared libraries/PD Client → Java Client → Hubble/Loader, and Python SDK → GraphMCP / AI applications
  • Validate downstream components with actual RC packages and record versions, image digests, and test results
• Validate the actual amd64/arm64 release images and manifests
  • Cover native library loading, startup, and core integration scenarios for Server, PD, Store, Hubble, and Loader according to the release scope
  • Limit RISC-V release claims to verified RocksDB-only support and back any performance multipliers with reproducible data

Final 1.8.0 integration acceptance

These checks summarize completion of the repository-specific work above

• Confirm release scope and candidate-feature decisions, with all required repository TODOs completed or explicitly deferred
• Pass deployment, upgrade, cross-repository integration, and supported-platform acceptance on the same RC set, with results traceable to the artifacts
• Close the release-specific documentation work in Doc #491, keeping release notes and support boundaries consistent with actual artifacts

SeaTunnel integration and documentation are complete; only its official 3.0 release is awaited, with voting already underway

Release procedure

Refer to the official website for Apache release validation and voting replies and email templates

• Prepare: assign the Release Manager, create the release branch, version updates, and RC tag from confirmed mainline commits, and check signing keys and publishing configuration
• Package and stage: generate source archives and the selected binary, Maven/Python, and image artifacts in repository dependency order, then stage the candidates
  • Build cleanly from the frozen RC, verify a common source revision for all artifacts, and record version and dependency mappings
• Validate candidates: complete the integration checks above and verify source builds, LICENSE/NOTICE, GPG signatures, SHA512 checksums, staging contents, and download links
• Community vote: submit candidate artifacts and validation records, complete voting and confirm the result; generate and validate a new RC if artifacts need changes
• Publish: release the approved artifacts, verify downloads, package registries, images, and the website, then announce the release