How to fix Error 10673 on Windows Server 2025 / Windows 11

by

Troubleshooting Guide: Error 10673 – Cannot Resume Index Creation Due to Incorrect Persisted State

Why This Error Happens

Error 10673 occurs when SQL Server 2025 attempts to resume a paused online index creation or rebuild operation, but the persisted state metadata in the RESUMABLE_INDEX internal tracking structures does not match the current state of the index or the underlying table schema. This typically happens when the table schema was altered while the index operation was paused, the temporary statistics or state information became corrupted due to an unexpected service interruption, or the index build was paused across incompatible SQL Server builds or patch versions. Because the resume operation relies on consistent persisted checkpoints, any mismatch forces SQL Server to reject the resume request and demand a full abort instead of risking a corrupted index.

Step-by-Step Resolution Guide

Step 1: Abort the Resumable Index Operation

  • Connect to the affected instance via SQL Server Management Studio (SSMS) or sqlcmd.
  • Run the following command to safely terminate the stuck operation:
ALTER INDEX [YourIndexName] ON [Schema].[TableName] ABORT;
  • Confirm the operation is cleared by querying sys.index_resumable_operations to ensure no residual entries remain for that index.

Step 2: Validate Schema and Table Consistency

  • Check whether any ALTER TABLE or ALTER INDEX DDL statements were executed on the table while the index build was paused.
  • Run DBCC CHECKTABLE('TableName') WITH NO_INFOMSGS, ALL_ERRORMSGS; to confirm there is no underlying page-level or allocation corruption contributing to the mismatch.
  • If corruption is found, address it using standard DBCC CHECKDB repair procedures before proceeding.

Step 3: Clear Cached Plan and Temporary State Data

  • Clear the procedure cache for the affected database to eliminate any stale execution plans referencing the broken resumable state:
DBCC FREEPROCCACHE;
  • Restart the SQL Server service via Windows Server 2025 Services Manager (services.msc) or PowerShell:
Restart-Service -Name 'MSSQLSERVER' -Force
  • This ensures in-memory resumable index metadata is fully flushed rather than relying on a stale cached state.

Step 4: Recreate the Index From Scratch

  • Once the operation is aborted and the cache is cleared, rebuild the index as a fresh, non-resumed operation:
CREATE INDEX [YourIndexName] ON [Schema].[TableName] ([ColumnName])
WITH (ONLINE = ON, RESUMABLE = ON, MAXDOP = 4);
  • Monitor progress using sys.dm_exec_requests and percent_complete to confirm the build is progressing normally without state errors.

Alternative or Verification Steps

  • Verify the fix by re-running SELECT * FROM sys.index_resumable_operations; — the table should show the index state as either actively running or completed, with no “paused” entries referencing corrupted state.
  • If the error persists after a clean abort and rebuild, check for mixed SQL Server patch levels in an Always On Availability Group or cluster setup, as resumable index state is not guaranteed to be compatible across mismatched Cumulative Update (CU) versions — apply the latest CU uniformly across all nodes and retry.