Wednesday, September 17, 2014

Databases are suffering from "Checkpoint not complete" in alertlog


V$log listing all the groups are in  ACTIVE and CURRENT status and alert log listing checkpoint not complete

Setting the parameter archive_lag_target   from 180sec to 900 sec resolved the issue, 

Check the impact of setting archive_lag_target in your environment

datafilecopy header validation failure for file +DATA due to file move from file system to ASM via sqlplus

datafilecopy header validation failure for file +DATA

This issue happens due to file move from file system to ASM via sqlplus

The renaming of the datafiles was done via SQL Plus and NOT via RMAN with SWITCH command

Error message looks like "datafilecopy header validation failure for file +DATA"

Solution :

During  RMAN COPY command to move datafiles to a new location (ASM or otherwise), do NOT rename the file via SQLPlus but use the RMAN SWITCH command instead (SWITCH DATAFILE ALL).

Check the alertlog for the message
alter database rename file '/u01user01.dbf  to  '+DATA’;


If SQLPlus has already been used to rename the files the rman metadata can be corrected as follows:

1. Check name of the current datafiles:

RMAN>report schema;

2. Confirm the name of the datafile copies (this will show the SAME list of datafiles):

RMAN>list copy of database;

3. Remove the datafilecopy entries from the rman repository (this will NOT remove the physical
files but simply update the rman repository):

RMAN>change datafilecopy uncatalog;

You can get value from the list copy of database output - do this for every datafile
that is 'duplicated' in the list copy output.

4. Confirm that the copies have been removed:

RMAN>list copy of database;