TIKA-4606: Upgrade Apache Ignite from 2.x to 3.x #2505
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Upgraded ignite.version from 2.17.0 to 3.1.0 - Replaced Ignite 2.x dependencies with Ignite 3.x equivalents: - ignite-core → ignite-api + ignite-runner - ignite-spring → removed (not needed) - Removed H2 database dependency (Calcite is built-in to Ignite 3.x) - Added exclusions for REST and metrics modules (not needed for config store) - Added dependency management to resolve convergence issues: - kotlin-stdlib: 2.2.0 - picocli: 4.7.5 - micronaut-inject: 3.10.4 - snakeyaml: 2.4 ✅ Calcite SQL engine now built-in via ignite-sql-engine ✅ No H2 dependency ❌ Code refactoring still needed - compilation errors due to API changes (Ignite 2.x cache API → Ignite 3.x table API) Next: Refactor IgniteConfigStore, IgniteStoreServer, IgniteConfigStoreConfig to use new Ignite 3.x Table API and configuration
✅ COMPILATION SUCCESS - All code refactored for Ignite 3.x API Changes: 1. IgniteConfigStoreConfig.java: - Replaced CacheMode enum with replicas/partitions - tableName replaces cacheName (Ignite 3.x uses tables not caches) - Added partitions configuration - Removed getCacheModeEnum() method 2. IgniteConfigStore.java: - Complete rewrite for Ignite 3.x client-server architecture - Uses IgniteClient.builder() to connect to cluster - KeyValueView<K,V> replaces IgniteCache<K,V> - Table-based storage instead of cache-based - Client-server model (connects to IgniteStoreServer) 3. IgniteStoreServer.java: - Uses IgniteServer for embedded server - Creates tables and distribution zones via SQL - Simplified initialization (no complex config needed) - Uses Ignite 3.x Table API 4. IgniteConfigStoreTest.java: - Updated to use BeforeAll/AfterAll for server lifecycle - Starts IgniteStoreServer once for all tests - Clients connect to server instance Technical Details: - Client connects via port 10800 (default) - Distribution zones configure replication - SQL: CREATE ZONE, CREATE TABLE - KeyValueView for simple get/put operations - SQL queries for keySet() and size() Status: ✅ Code compiles successfully ✅ No dependency issues ✅ Checkstyle passes ✅ Spotless passes⚠️ Tests need server initialization fix (Ignite 3.x embedded startup) Next: Fix embedded Ignite 3.x server startup in tests
… 3.x Changes: 1. tika-parent/pom.xml - Added dependency management for Ignite 3.x convergence: - org.ow2.asm:asm:9.9.1 (was conflicting 9.9 vs 9.9.1) - info.picocli:picocli:4.7.7 (was conflicting 4.7.5 vs 4.7.7) - org.yaml:snakeyaml:2.4 (was conflicting 2.0 vs 2.4) - javax.validation:validation-api:2.0.1.Final 2. TikaGrpcServerImpl.java - Updated startIgniteServer() for Ignite 3.x: - Replaced CacheMode with replicas/partitions - tableName instead of cacheName (backwards compatible) - Uses new IgniteStoreServer(tableName, replicas, partitions, instanceName) - Parses both old (cacheName) and new (tableName) config for compatibility Result: ✅ BUILD SUCCESS with no convergence errors
- Upgraded ignite-api, ignite-client, and ignite-runner to 3.1.0 - Migrated from cache-based to table-based API - Updated configuration to use tableName instead of cacheName - Added dependency management for Micronaut dependencies to resolve convergence issues - Updated forbidden API calls to use Locale.ROOT - Modified IgniteStoreServer to use Ignite 3.x API and configuration - Build succeeds and basic gRPC tests pass - Ignite 3.x runtime requires further investigation for proper server startup
- Upgraded ignite-core 2.16.0 -> ignite-runner 3.1.0 - Migrated from IgniteConfiguration to hocon-based config - Updated IgniteConfigStore to use new KeyValueView API - Fixed IgniteStoreServer for embedded mode - Updated ExtensionConfigDTO to use Ignite 3 Mapper - Added required JVM --add-opens flags for Java 17+ - Fixed EmitHandler NPE for NO_EMIT scenario - Added emitter_id to FetchAndParseRequest proto - Integrated e2e tests into parent build - Added local server mode for CI (no Docker required) - Fixed gRPC channel resource leak in tests - All 11 unit tests passing, e2e test passing
Contributor
Author
|
CI workflows encountered a transient Maven Central 403 Forbidden error (not related to this PR's changes). All workflows have been re-run and are now executing successfully. This is a known intermittent issue with GitHub Actions and Maven Central repository access. |
Contributor
Author
|
Fixed Windows build failure 🪟 The Windows CI was failing with: Root cause: Ignite 3.x adds many more dependencies than 2.x, causing the classpath to exceed Windows command line limits (~8191 characters). Solution: Implemented Java @argfile support in PipesClient:
This is a general improvement that will help any future dependency additions. ✅ |
Contributor
Author
|
Fixed forbiddenapis check ✅ The Fix: Changed to All builds should now pass! 🚀 |
nddipiazza
force-pushed
the
TIKA-4606-ignite-3x-upgrade
branch
from
ff538a9 to
6f2462e
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
This PR upgrades Apache Ignite from 2.16.0 to 3.1.0 in the tika-pipes-config-store-ignite module.
Changes Made
Core Upgrade
Server & Integration
Testing & CI
Test Results
✅ 11/11 unit tests passing in tika-pipes-config-store-ignite
✅ E2E test passing - processes documents successfully
✅ No resource leaks - proper cleanup verified
✅ BUILD SUCCESS locally
Breaking Changes
None - API remains backward compatible from user perspective
CI Configuration
Tests use local server mode by default:
tika.e2e.useLocalServer=true-Dtika.e2e.useLocalServer=falseto use DockerFixes apache/tika#TIKA-4606