3078958 - Indexserver slow startup because long recovery queue

• You observe the indexserver is taking longer to restart after one of the following scenarios:
  
  • Ungraceful shutdown of an SAP HANA system
  • Removal of a system from SAP HANA System Replication
• You observe the indexserver service remains in yellow state in SAP HANA Studio or the output of sapcontrol command
• In the indexserver trace you can observe progressing entries similar to the following:

[73447]{-1}[-1/-1] 2021-07-19 13:07:59.204597 i Logger           RecoveryHandlerImpl.cpp(00735) : Redo done up to position: 0x236fb68f40 and time: 2021-07-19 09:59:55.934557-07:00 (47%)
[72452]{-1}[-1/-1] 2021-07-19 13:11:59.156403 i PITRestart       LogReplayCoordinator.cpp(00611) : Log Replay Coordinator (master, volume=3)
 - lastReplayStepStartLogPos: 0x237296b5c9
Replay Step Information
 - replayStepStatus: ReplayStep_Running
[75132]{-1}[-1/-1] 2021-07-19 13:12:05.014295 i Logger           RecoveryHandlerImpl.cpp(01790) : Triggering log recovery up to position 0x2372d6b5c9

---

[361765]{-1}[-1/-1] 2022-08-17 20:49:33.415432 i CSARedo          RedoMerge.cpp(02747) : MergeInfoPersistenceManager instance for log replay created: instance @ 0x00007f6eabebb000, associated persistence manager @ 0x00007f7d223b3018
[361765]{-1}[-1/-1] 2022-08-17 20:49:33.415452 i CSARedo          RedoMerge.cpp(01303) : Merge and optimize compression configuration during log replay: active = 0, merge type = NO_MERGE, optimize compression type = NO_OPTIMIZE_COMPRESSION, cancel jobs before going online = 1, delta row limit for enforced merges = 500000000, delta row limit for enforced synchronous merges = 1610612736, timeout for index handle lock = 120000, ignored tables = 
[361765]{-1}[-1/-1] 2022-08-17 20:49:33.415458 i CSARedo          Configuration.cpp(00142) : redo replay settings: AllowCachingOfRedoDmlContext = 1, UseDmlExecutorBatchRedo = 1, UseDmlExecutorBatchSortRedo = 1, AllowMergeAndOcDuringRecoveryToFinishAnytime = 1, MaxRowsToLog = 0, UsePreferredNumaNodeForPreload = 1, MaxSizeToLog = 209715200, DisabledTableConsistencySubChecks = 12, WarmupperLoadDeltaFragments = 1, TreatFirstNonDdlRedoOfContainerAsNoOpenCch = 1, MaxLoadUnitConversionThreads = 8, OptimizeReplicaForOnlineDdlInLogReplay = 1, IgnoreDelta2DuringDdlRedo = 0, UseOperationModeLogReplayReadAccess = 0, FirstDataItemToTrace = 0, NumDataItemsToTrace = 10, ParallelWriteThreshold = 2000
[361771]{-1}[-1/9223372036854775806] 2022-08-17 20:49:35.617143 d AvcLifecycle     Configuration.cpp(00103) : AvcLifecycleConfig: getAllowEarlyFragmentCleanupAtMerge: 1, getAllowEarlyFragmentCleanupAtUnload: 1, getMaxMergedMainFragmentsPerTablePart: 0, getAlwaysFuseIntoNewDeltaFragment: 1
[361261]{-1}[-1/-1] 2022-08-17 20:59:41.172089 i Logger           RecoveryHandlerImpl.cpp(00759) : Redo done up to position: 0x31579d6e00 and time: 2022-08-16 17:41:47.244286+09:00 (0%)
[361670]{-1}[-1/-1] 2022-08-17 21:09:51.181568 i Logger           RecoveryHandlerImpl.cpp(00759) : Redo done up to position: 0x31579d6e00 and time: 2022-08-16 17:41:47.244286+09:00 (0%)

"Image/data in this KBA is from SAP internal systems, sample data, or demo systems. Any resemblance to real data is purely coincidental."

SAP HANA, platform edition 2.0

yellow state, index server, not starting, started , KBA , HAN-DB-PER , SAP HANA Database Persistence , Problem