This topic describes frequently asked questions about the backup and recovery features of the OCP system.
Supported features
Q1: Is OCP compatible with the backup and recovery feature of OSS ?
A: OCP supports OSS-based logical backup and recovery since OCP V2.2.x and OSS-based physical backup and recovery since OCP V2.5.1.
Storage configuration
Q1: Why did the storage address test fail during backup and recovery ?
A: The admin user does not have the write privilege on the NFS directory.
Q2: The NFS file system is mounted on the OCP server and the three OceanBase servers. Why did storage configuration detection fail on the three OBServers ?
A: Maybe you do not have the write privilege on the /obbackup directories on these three OceanBase servers.
Q3: The system indicates that the storage address test failed when I initiate recovery in the global view. What can I do?
A: In OCP V2.5.3 and later versions, when a recovery is initiated in the global view, the system tests whether the storage address has read and write permissions on the selected host. We recommend that you use the ls -l command on the selected host to check the read and write permissions on the storage directory:
rwx indicates that the user of the file can read, write, and execute the file.
r-x indicates that the user group of the file can read, but not write or execute the file.
r-x indicates that other users can read and execute, but not write the file.
Backup
Q1: Why is the error Cluster compaction needed before full data backup returned in the pre-check step when I use OCP to execute a full backup ?
A: You must perform a major freeze on clusters before performing a full backup. For more information, see Perform a major compaction.
Q2: In the Pre-check before data backup step of a backup task, an error is returned, saying data backup pre-check failed, full data backup needed before incremental data backup . Why ?
A: A full backup must be implemented before incremental backup.
Q3: I initiated a full data backup task in OCP. Then, the end time is displayed in the Ended At column. Does that mean that the backup task is completed?
A: Check the Status column to find out whether a task is completed.
Recovery
Q1: Can I restore the data in a tenant in a full backup to an existing tenant in the same cluster?
A: The data can be restored to the same cluster but it must be restored to a new tenant. A new tenant is automatically created during the restoration process.
Q2: I have initiated a data restoration task, and I saw that the backend resource pool has been created, but the tenant was not created. Why?
A: This is because no NFS has been mounted to the destination cluster.
Q3: During physical recovery, the system reported that restore_concurrency not set . What is the cause of this problem?
A: Generally, it may take some time for the restore_concurrency parameter of OceanBase Database to take effect, and OCP will usually return an error. You may ignore the error after you commit Initiate Recovery without any error.
Q4: Do I need to back up MetaDB first before I recover the OCP and OceanBase clusters after they are cleared ?
A: No. You can directly restore the data by selecting the backup file on the new server where OCP is installed. In the NFS scenario, you must mount the backup directory to the destination cluster for recovery.
Q5: When I recover my OceanBase cluster , a log error occurred. Error code [9023]; cannot start log archive backup; . What is the cause of this problem?
A: You must avoid running run the command repeatedly to enable log backup.
Inspection
Q1: When I was using the CLI tool to back up and recover clusters in OceanBase Database V2.2.7x, an error is returned during backup and recovery inspection. What is the cause of this problem?
A: This usually happens when an exception was thrown during backup and recovery and you ignored it. You can set the valid field to 0 for backup and recovery inspection in the ocp_inspect_task table to disable inspection.