RMAN: "WARNING: nn objects could not be deleted ..." and backup takes too long
- Article Type: General
- Product: Aleph
- Product Version: 20, 21, 22, 23
Problem Symptoms:
Running the Ex Libris Backup Package ora_cold procedure (in /exlibris/backup/scripts/exec_backup_main), the backup seems to be successful, but takes extremely long (6 hours for a 300,000-bib-record database).
The /exlibris/backup/logs/ora_cold_a1_Detail_nnnnnn_nnnn.log shows:
ora_cold_a1_Detail_140130_2048.log:Deleted 10 objects
ora_cold_a1_Detail_140130_2048.log:RMAN-06207: WARNING: 16 objects could not be deleted for DISK channel(s) due
ora_cold_a1_Detail_140130_2048.log:RMAN-06210: List of Mismatched objects
It's unclear just how these objects which could not be deleted differ from those which were deleted.
Cause:
In most cases the hot or cold backup taking so much longer than it should is due to the presence of a large number of Oracle archive log files in the /arch directory. The "could not be deleted" messages confirm this in this case.
Resolution:
If it's clear that the cold backup was successful (or if there's another cold backup which could be used, for instance, a successful Oracle Data Pump backup -- see the article " Refreshing Test server data Using Oracle Data Pump " --), do the following:
* (temporarily) turn off archiving (util o/7/2)**.
* rename the /arch directory to /arch.old
* do "mkdir arch" to recreate it (empty)
* do util o/7/1** to restart archiving
** Note: The util_o_07_02 and util_o_07_01 procedures both stop and restart Aleph. They should be run only when the system is not in use.
Additional Information
1. The $aleph_proc/util_o_07_02 file has the following lines:
...
alter system switch logfile;
...
alter database noarchivelog;
and the $aleph_proc/util_o_07_01, these lines:
...
alter system switch logfile;
...
alter database archivelog;
2. The following RMAN commands may be of use:
? delete archivelog all
? crosscheck archivelog all
? delete expired archivelog all
- Article last edited: 5/16/2014