Applies toOCI Base Database Service (VM DB systems), Oracle Database 19c/23ai, RMAN
See the failure
oci db backup list --database-id $DB \
--query 'data[].{name:"display-name",type:type,state:"lifecycle-state",start:"time-started",details:"lifecycle-details"}' --output table
Also check Work Requests on the database and the DB system for the error text.
Connectivity to Object Storage
Automatic backups go to Oracle-managed Object Storage.
- The DB system subnet has a route to Object Storage: a service gateway with "All <region> Services" (or the Object Storage label), or a NAT gateway.
- Security rules allow egress HTTPS (443) to that service CIDR label.
- No host proxy or firewall change blocks
objectstorage.<region>.oraclecloud.com.
# from the DB node
curl -s -o /dev/null -w "%{http_code}\n" https://objectstorage.ap-mumbai-1.oraclecloud.com
Space on the database host
- The FRA isn't full. Archive logs pile up if backups fail, and that makes everything worse.
SELECT name, ROUND(space_limit/1024/1024/1024) limit_gb, ROUND(space_used/1024/1024/1024) used_gb, ROUND(space_reclaimable/1024/1024/1024) reclaimable_gb FROM v$recovery_file_dest; SELECT file_type, percent_space_used FROM v$recovery_area_usage; - Local filesystems used by the tooling (
/u01,/var) have free space:df -h.
RMAN health
sudo su - oracle
rman target /
RMAN> SHOW ALL;
RMAN> LIST BACKUP SUMMARY;
RMAN> CROSSCHECK ARCHIVELOG ALL;
RMAN> REPORT NEED BACKUP;
| Error | Meaning |
|---|---|
ORA-19809: limit exceeded for recovery files | FRA full. Back up and delete archived logs, or grow db_recovery_file_dest_size temporarily. |
ORA-19511 with media manager messages | Object Storage module / connectivity problem |
RMAN-06059: expected archived log not found | Archive logs deleted outside RMAN. CROSSCHECK and DELETE EXPIRED, then take a fresh full (level 0) backup. |
ORA-01031 / wallet errors | Tooling wallet or credentials changed. Don't edit tooling-managed configuration by hand. |
Tooling
- Cloud tooling on the DB system is up to date (see the console's Updates for the DB system).
- Nobody changed the RMAN configuration (channels, retention) that tooling manages. Custom changes can break automatic backups.
After the fix
- Trigger an on-demand full backup from the console, so you have a clean base.
- Check that the next scheduled incremental succeeds.
- Make sure archive log backups run and the FRA usage drops.
Important
Test a restore to a new DB system or a point in time every quarter. A successful backup job alone doesn't prove you can recover.