When your DC replication has halted, check whether there's an entry in the event log: NTFRS event id 13568 (JRNL_WRAP_ERROR).

This error occurs when the DC on which this event id is generated has been powered off unexpectedly, or when the system volume was full. The error signifies that the database journal (used during replication) has been cut off, resulting in the replication being halted, which is by design.

To resolve, you can tell NtFRS to automatically repair the problem and continue with the replication by changing the value of a certain registry parameter. The parameter can be found at HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\NtFrs\Parameters\Backup/Restore\Process at Startup. There, you'll find a DWORD value called BurFlags. Set the value to:

• D4, performs an authoritative mode restore;
• D2, performs a non-authoritative mode restore.

Usually, you run a authoritative mode restore on a PDC, but in case this fails, try it with a non-authoritative mode restore.

Before making the change, first stop NtFrs. After the change, don't forget to restart your DC, or, simply start NtFrs again.