[BUG]Redis 7.4.5 aof continuous backup failed

[halcyon-r]

Describe the bug
When I deploy a replication cluster for Redis 7.4.5, and configured aof continuous backup. I found errors and the aof backup file was not uploaded correctly.

To Reproduce
Steps to reproduce the behavior:
After deploy kubeblocks v1.0.1 and redis-addon 1.0.1.

1. create a redis replication cluster with version 7.4.5:

apiVersion: apps.kubeblocks.io/v1
kind: Cluster
metadata:
  name: redis-7
  namespace: test-redis
spec:
  terminationPolicy: Delete
  clusterDef: redis
  topology: replication
  services:
  - componentSelector: redis
    name: redis-7-internet
    serviceName: redis-7-internet
    roleSelector: primary
    spec:
      ipFamilyPolicy: PreferDualStack
      ports:
        - name: redis
          port: 6379
          protocol: TCP
          targetPort: redis
      type: LoadBalancer
  componentSpecs:
    - name: redis
      env:
      - name: REDIS_DEFAULT_PASSWORD
        value: ""
      serviceVersion: "7.4.5"
      disableExporter: false
      replicas: 2
        #services:
        #- name: redis-advertised
        #serviceType: LoadBalancer
        #podService: true
      resources:
        limits:
          cpu: '0.5'
          memory: 1Gi
        requests:
          cpu: '0.5'
          memory: 1Gi
      volumeClaimTemplates:
        - name: data
          spec:
            storageClassName: ""
            accessModes:
              - ReadWriteOnce
            resources:
              requests:
                storage: 2Gi
    - name: redis-sentinel
      serviceVersion: "7.4.5"
      disableExporter: false
      replicas: 3
      resources:
        limits:
          cpu: '0.5'
          memory: 0.5Gi
        requests:
          cpu: '0.5'
          memory: 0.5Gi
      volumeClaimTemplates:
        - name: data
          spec:
            storageClassName: ""
            accessModes:
              - ReadWriteOnce
            resources:
              requests:
                storage: 2Gi

2. Then configure BackupRepo and BackupSchedule to start aof incremental backup.

$ ku get backup -n test-redis
NAME                                       POLICY                        METHOD            REPO                 STATUS      TOTAL-SIZE   DURATION   DELETION-POLICY   CREATION-TIME          COMPLETION-TIME        EXPIRATION-TIME
ff67ea12-redis-7-redis-backup-aof          redis-7-redis-backup-policy   aof               s3-compatible-repo   Running     1160                    Delete            2025-10-11T11:10:04Z            

3. Observe that the statefulset for aof has been successfully created and running.

$ kubectl get po -n test-redis
NAME                                    READY   STATUS             RESTARTS         AGE
ff67ea12-redis-7-redis-backup-aof-0     2/2     Running            0                74m

4. But the log of the job pod shows some errors.

$ kubectl 
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected
2025-10-11 12:10:03 INFO: Update file: /data/appendonlydir/appendonly.aof.2.incr.aof
2025-10-11 12:10:04 INFO: save_backup_status: empty start_time from backup repo, use current time
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected
environment: line 201: [: -gt: unary operator expected

And the "start" and "end" field for the Backup are always the same.

$ kubectl get backup -n test-redis ff67ea12-redis-7-redis-backup-aof  -oyaml | yq '.status.timeRange'
{
  "end": "2025-10-11T12:10:04Z",
  "start": "2025-10-11T12:10:04Z"
}
$ ku get backup -n test-redis ff67ea12-redis-7-redis-backup-aof  -oyaml | yq '.status.timeRange'
{
  "end": "2025-10-11T12:10:37Z",
  "start": "2025-10-11T12:10:37Z"
}
$ ku get backup -n test-redis ff67ea12-redis-7-redis-backup-aof  -oyaml | yq '.status.timeRange'
{
  "end": "2025-10-11T12:10:45Z",
  "start": "2025-10-11T12:10:45Z"
}

Check the S3 bucket, the aof file was not uploaded as expected.

Expected behavior
The "start" and "end" field of Backup should not always be the same, and aof files should be uploaded correctly.

Screenshots

Desktop (please complete the following information):

• Version: Kubeblocks v1.0.1，Redis Addon v1.0.1

Additional context
I tried to debug the script pitr-backup.sh and found that the redis tools version is 7.2.10，which cannot match the Redis version 7.4.5. This makes the get_base_file_ctime() method unable to get the correct result, which leads to a "-gt: unary operator expected" problem. The if conditional statement was not executed and the aof file is not uploaded.

if [ $(get_base_file_ctime "$base_file") -gt ${global_aof_last_modify_time} ]; then
  datasafed push "${base_file}" "${backup_files_prefix}.dir/${AOF_DIR}/$(basename "${base_file}")"
  datasafed push "${backup_manifest}" "${backup_files_prefix}.dir/${backup_manifest}"
  datasafed push "${DATA_DIR}/users.acl" "${backup_files_prefix}.dir/users.acl"
  DP_log "Upload file: $base_file $backup_manifest users.acl"
fi

Here is the redis-check-rdb tool error log:

root@redis-backup-aof-test-0:/# redis-check-rdb /data/appendonlydir/appendonly.aof.2.base.rdb
[offset 0] Checking RDB file /data/appendonlydir/appendonly.aof.2.base.rdb
--- RDB ERROR DETECTED ---
[offset 9] Can't handle RDB format version 12
[additional info] While doing: start
[additional info] Reading type 0 (string)
[info] 0 keys read
[info] 0 expires
[info] 0 already expired

I have checked the redis addon v1.0.2. and I think this problem has not been resolved in redis addon v1.0.2.

[shanshanying]

@wangyelei pls help check if the error exists on redis addon 1.0.1 and 1.0.2