Special Scenario - Loss of Datafile (non-critical) without any backup and in Noarchivelog mode

 We faced below interesting issue in one of our test environments, I call it SPICE19 here. One of the datafile was deleted with the intent to recreate it. Unfortunately, database was up and running at that moment when the datafile was deleted at OS level.

Observations -

1. Database SPICE19 was in noarchivelog mode.

2. No backups were available for restoration.

3. Datafile '/u02/oracle/oradata/SPICE19/users06' was created accidentally with a wrong name (should have been /u02/oracle/oradata/SPICE19/users06.dbf)

4. It was then removed using (rm) at OS level.

5. A new datafile was added with the name - /u02/oracle/oradata/SPICE19/users06.dbf

6. Database got crashed searching for datafile - /u02/oracle/oradata/SPICE19/users06

7. Database complained for missing datafile - /u02/oracle/oradata/SPICE19/users06 as the information was in controlfile.

 

Troubleshooting Steps -

1. Missing datafile id was found below -

SQL>  select * from v$recover_file;

     FILE# ONLINE  ONLINE_ ERROR                              CHANGE# TIME          CON_ID

---------- ------- ------- ------------------------------------------ --------- ----------

        18 ONLINE  ONLINE  FILE NOT FOUND                           0                    0

 

2. Second step was to recreate missing datafile with a different name-

SQL> alter database create datafile '/u02/oracle/oradata/SPICE19/users06' as '/u02/oracle/oradata/SPICE19/users06_new.dbf' reuse;

Database altered.

 

3. As the database was in norachivelog mode, recovery is not possible UNLESS complete information is present in the REDO logs and they are not overwritten.

 

4. We have 3 redo logs here,

/u02/oracle/oradata/SPICE19/redo03.log

/u02/oracle/oradata/SPICE19/redo02.log

/u02/oracle/oradata/SPICE19/redo01.log

 

5. Tried applying them in the following sequence with incomplete recovery -

SQL> recover database using backup controlfile until cancel;

/u02/oracle/oradata/SPICE19/redo01.log

/u02/oracle/oradata/SPICE19/redo02.log

Log applied.

Media recovery complete.

 

6. Opened the database with resetlogs option

 

Correct Approach -

We must rename datafile with below 2 steps.

1. Tablespace must be offline (no users can access the data during this period)

alter tablespace USERS offline immediate;

 

2. Rename the datafile at OS level using mv command for Linux OS.

mv /u02/oracle/oradata/SPICE19/users06 /u02/oracle/oradata/SPICE19/users06.dbf

 

2. Rename the datafile using below command to update the controlfile at database level -

ALTER TABLESPACE users rename datafile '/u02/oracle/oradata/SPICE19/users06'

to '/u02/oracle/oradata/SPICE19/users06.dbf';

 

Conclusion-

We were lucky to have the redo information in the redo logs and they were not overwritten. This was a rare case where recovery of a datafile was possible with database in noarchivelog mode and with no backups available.