Version 1.1.0.0
Version 1.1.0.0 Release Notes
Compatible with OpenSearch 1.1.0.0
Features
- Porting replication plugin to OpenSearch (#8)
- Pause and Resume API (#10)
- Leader Stats API (#122)
- Follower stats API (#126)
- AutoFollow Stats API Addition (131)
Enhancements
- Configure the threadpool for leader cluster (#14)
- Renew retention lease with the last known synced checkpoint (#18)
- Handling node disconnections and shard movement (#23)
- Enforce metadata write block on follower indices (#120)
- Support for storing security metadata and Integration with security plugin (#15)
- Fix for race condition with snapshot restore (#28)
- Status api changes (#26)
- Changed names for assume roles to remote and local from leader and follower (#29)
- Added preference to fetch the metadata from the primary shard and additional logging for the metadata store (#29)
- Handling leader cluster disconnection failure scenarios during bootstrap (#17)
- Support for translog fetch at the leader(remote) cluster and enable translog pruning during replication setup (#25)
- Added random active shard selector for the getchanges operations (#33)
- Settings and Alias replication (#22)
- Rename opendistro endpoints (#36)
- Added connections alias and doc count in status api (#38)
- Bootstrapping bug fix and few integ tests. (#41)
- Adding validation on index settings (#37)
- Renamed connection alias and changed exception handling (#44)
- Added non verbose version of status api. (#46)
- Added replication specific settings for recovery (#46)
- Add support for parallel reader and writer for ShardReplicationTask (#48)
- Changes to fail chain replication (#50)
- Using non null RestoreInProgress value in case of no restore in progress (#54)
- Not allowing index starting with . for replication (#55)
- Changing the default recovery chunk size to 10MB (#58)
- Added reason to the replication metadata store (#57)
- Propagate shard failures to index task and auto-pause on failures (#56)
- Propgate 4xx failures from ShardReplicationTask (#60)
- Bug fixes and autofollow task resiliency (#59)
- Added permission validation steps at user level and role level (#62)
- Async metadata update. Separated metadata reader and updater (#52)
- Validate analyzer path when starting/resuming replication (#64)
- Prevent starting replication directly on the leader index alias. (#66)
- Change to stop showing replication related metrics in case of non running (syncing and bootstrap) state. (#68)
- Integ test for follower index deletion (#68)
- Adding reason while serializing PauseReplication Request (#71)
- Add integ test to verify mapping propagation (#72)
- Integ test to verify that delete/index operation are blocked on follower (#73)
- Changes to support security plugin integ tests (#70)
- Integ test: Verify that replication is paused when remote index is deleted (#74)
- added few integ tests when open and close is triggered on leader index (#77)
- Integ test to stop replication when remote cluster is unreachable (#69)
- Moved validation checks after setup checks (#85)
- Modified tests to use custom roles for default tests (#83)
- Replication engine tests: Index operations on leader (#79)
- Add Handling of IndexReplicationTask cancellation and corresponding ITs (#87)
- Updated security plugin with latest changes (#89)
- Integ test: forcemerge and snapshot on leader during bootstrap (#76)
- Validate that soft delete settings is enabled at leader (#84)
- Adding Security for replication APIs and IT for DLS, FLS or FieldMasking (#90)
- changed few naming conventions. 1. remote to leader, 2. local to follower, etc (#75)
- Refactor code for OpenSearch (#98)
- Port terminology from OpenDistroForElasticsearch to OpenSearch. (#98)
- Changed the dependency to OS 1.1 (#100)
- Rename the replication settings (#103)
- Support for mapping update for the replication store system index (#105)
- Adding UTs for IndexReplicationTask (#109)
- OpenSearch Integ tests for reroutes (#114)
- Adding exp backoff for replay and get changes (#135)
- Rename to useroles in all the request payloads instead of assumeroles (#154)
- Initializing shard metric for UT (#141)
- Changing the log level to debug as this information is surfaced in stats (#149)
- Pause API should allow reason to be specified in the REST call. (#151)
- Verifying shard tasks are up in autofollow test (#153)
- Correcting param name for leader API (#163)
- Populate "AutoPaused: " prefix in the pause-reason while auto-pausing. (#170)
Bug Fixes
- Fix submitClusterStateUpdateTask source (#6)
- Throw exception if replication is not enabled on the remote domain (#86)
- Increase wait time for snapshot/forcemerge integ tests (#91)
- Add validation on index name (#92)
- Remove opendistro security zip (#95)
- Changes to restore default threadcontext after co-routine is suspended (#102)
- Remove packaged example as replication plugin will ship with OpenSearch. (#107)
- Handled errors during test role update (#112)
- Ignore forbidden APIs to access private members of ES datastructure (#120)
- Starting missing shard tasks (#114)
- Fix for RetentionLeaseInvalidRetainingSeqNoException while adding retention lease (#128)
- Fix IT Test (#137)
- Fix RetentionLeaseNotFoundException during shard reroute on follower (#138)
- Fix Flaky integ test (#140)
- Handled rename change for useroles in tests (#142)
- Remove the flaky assert in StartReplicationIT (#144)
- Blocked k-NN index for replication (#158)
- Gracefully fail replication on bootstrap failure. (#116)
Infrastructure
- Add OpenSearch build script (#119)
- Add basic github workflow for build and test (#123)
- Modified github workflow to pick the latest changes (#173)