Creating a new Pluggable Database (PDB) on a primary database is usually a routine task. However, in Oracle AI Database 26ai, DBAs often face an unexpected issue where Data Guard recovery on the physical standby gets stuck immediately after a new PDB is created on the source.
This article walks through a real production scenario, explains why Data Guard breaks, and shows the exact steps to fix ORA-19723 and UNNAMED datafile issues safely — without rebuilding the standby.
Problem Overview
After creating a new PDB (SALSE_PDB) on the primary database, managed recovery on the physical standby (teststby) stopped and began reporting errors similar to the following:
ORA-19723: cannot re-create plugged in data file 13
ORA-01111: name for data file 13 is unknown
ORA-01110: data file 13: '/u01/app/oracle/product/23.26.0.0/dbhome_1/dbs/UNNAMED00013'
ORA-27037: unable to obtain file status
Linux Error: No such file or directory
At this point:
- Redo transport is working
- Managed recovery is stuck
- The standby database cannot apply changes for the new PDB

Why This Happens
This issue occurs due to a timing and configuration mismatch during PDB creation.
Root Causes
- New PDB created on primary
- New datafiles generated for the PDB
- Standby attempts to create matching datafiles
standby_file_managementis not set to AUTO- Oracle creates a placeholder file:
UNNAMED00013 - Recovery halts because Oracle cannot map the datafile correctly
This is not a bug — it’s expected behavior when standby automation is incomplete.
Symptoms You’ll See
- Data Guard configuration shows DISABLED
- Managed recovery stops silently
- Standby alert log reports
UNNAMEDdatafiles - PDB remains in MOUNTED state on standby
ALTER DATABASE CREATE DATAFILE ... AS NEWfails with ORA-19723
Failed Attempt: Manual Datafile Creation
A common first reaction is to manually create the datafile:
alter database create datafile
'/u01/app/oracle/product/23.26.0.0/dbhome_1/dbs/UNNAMED00013'
as NEW;
❌ This fails with:
ORA-19723: cannot re-create plugged in data file
Why?
Because this datafile belongs to a plugged-in PDB, and Oracle does not allow manual recreation in this scenario.
Correct Fix: Step-by-Step Solution
Step 1: Check PDB Status on Standby
show pdbs;
Output shows the new PDB stuck in MOUNTED mode:
CON_ID CON_NAME OPEN MODE
4 SALSE_PDB MOUNTED
Step 2: Switch to the Affected PDB
alter session set container=SALSE_PDB;
Step 3: Offline and Drop the UNNAMED Datafile
This tells Oracle to forget the broken placeholder.
alter database datafile '/u01/app/oracle/product/23.26.0.0/dbhome_1/dbs/UNNAMED00013' offline drop;
✅ This command succeeds and clears the blockage.
Step 4: Enable Automatic Standby File Management
This is the most critical fix.
alter system set standby_file_management='AUTO';
Without this parameter, every new PDB or datafile can break Data Guard again.
Data Guard Broker State Check
At this point, the broker may show the configuration as DISABLED.
dgmgrl /
show configuration;
Output:
DGMGRL> show configuration
Configuration - PROD to STBY
Protection Mode: MaxPerformance
Members:
test - Primary database
teststby - Physical standby database
Fast-Start Failover: Disabled
Configuration Status:
DISABLED
Re-enable the configuration
enable configuration;
Restart Managed Recovery
Oracle automatically resumes recovery once the file issue is resolved:
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE DISCONNECT NODELAY
You’ll see logs like:
- Parallel media recovery started
- Standby redo logs applied
- New PDB datafiles created automatically in ASM
- TEMP tablespaces dropped and recreated
- UNDO and USERS tablespaces synchronized
Confirmation in Alert Log
Successful recovery messages include:
SALSE_PDB(4): Successfully added datafile 14
SALSE_PDB(4): Successfully added datafile 15
SALSE_PDB(4): Successfully added datafile 16
Recovery successfully copied files for SYSTEM, SYSAUX, UNDOTBS1
This confirms:
- The new PDB is fully synchronized
- Data Guard is healthy
- No rebuild is required
Best Practices to Prevent This Issue
Always Set This on Standby
standby_file_management='AUTO'
Before Creating New PDBs on Primary
- Ensure managed recovery is running
- Verify sufficient space in ASM / FRA
- Monitor standby alert log during PDB creation
After Creating a PDB
- Check
show pdbson standby - Confirm PDB opens automatically after recovery
Key Takeaways
- UNNAMED datafiles after PDB creation are configuration issues, not corruption
- Never manually recreate PDB datafiles
- Dropping the UNNAMED file and letting recovery re-create it is the correct approach
- Oracle AI Database 26ai handles this cleanly when properly configured
- No Data Guard rebuild is needed if handled correctly
Final Thoughts
Oracle AI Database 26ai Data Guard is extremely resilient, but PDB operations require correct standby automation. A single missing parameter can stop recovery — yet the fix is simple once you understand what’s happening internally.
This scenario is a perfect example of why real-world DBA experience matters more than theory.




