🐞 bug: Each time PITR backup, the gs_probackup init operation is performed, and the agent reports an error.

[peilinqian]

Environment

• Contents of your values.yaml file,if your helm deployment process goes wrong.
• Contents of your CRD resource file. Include proxy.shardingsphere.apache.org/v1alpha1 and proxyconfig.shardingsphere.apache.org/v1alpha1
• Kubernetes distro (e.g. RKE/K3s/EKS/OpenShift) and version

Issue description

• Which phase is not functioning properly, is it the installation phase or the apply CRD phase ?
  PITR

Error log

• Including ShardingSphere-Proxy error log and ShardingSphere-Operator error log

{"level":"error","ts":"2023-12-26T10:36:03+08:00","caller":"pkg/opengauss.go:196","msg":"add instance failure[output=], err: exec failure[ee=exit status 1,stdout=ERROR: Instance 'ins-default-ss' backup directory already exists: '/home/peilq_og_sub/basebackup/backups/ins-default-ss'\n], wrap:10003-Command operate failed., wrap: 10006-The instance already exist."}

Expected result

The gs_probackup init operation is performed for each PITR backup. The agent displays a message but does not report an error.

[mlycore]

The error was logged and returned, and this may cause backup execution interruption in previous version.

And this is fixed as below:

• gs_probackup init will be executed by pitr-agent before any backup progress starts. feat: add Init before AddInstance while backing up #480
• And the backup data directory will also be created if not exist. fix: create backup path before checking disk #482