Skip to content

Sentinel mode to Sentinel mode

Redis-shake supports data synchronization and migration capabilities between Sentinel mode instances. In this example, we will demonstrate the configuration method for cross-cluster synchronization using a 3-replica Sentinel mode Redis.

Let's assume that instance redis-a and instance redis-b are in different clusters. Both instances are running in a 3-replica Sentinel mode. We will set up a synchronization structure where instance redis-a acts as the master instance and instance redis-b acts as the slave instance, providing the following disaster recovery support:

  • Under normal circumstances, instance redis-a serves external requests and continuously synchronizes data to instance redis-b.
  • If the master instance redis-a goes offline due to a failure, instance redis-b takes over and serves external requests.
  • Once the master instance redis-a recovers, the incremental data is written back from instance redis-b to instance redis-a.
  • After the data recovery of instance redis-a is complete, it switches back to the initial state where instance redis-a serves external requests and continues data synchronization with instance redis-b.

Note

Data synchronization and data recovery need to be performed by different Redis-shake instances. Therefore, in this example, we need to create two Redis-shake instances: Redis-shake-sync and Redis-shake-recovery .

Data Synchronization Deployment

Diagram: Data synchronization instance redis-a >> instance redis-b

Configuring Service for the Source Instance

If the source instance is in a DCE 5.0 cluster, you can enable the solution in Data Services - Redis - Cross-Cluster Master-Slave Synchronization , and the service configuration will be automatically completed.

If the source instance is in a third-party cluster, you need to manually configure the service. The configuration method is described below:

Create a NodePort service for the Redis instance to allow data synchronization access from Redis-Shake. In this example, we need to create one service for instance redis-a. Here's the process:

  1. Go to Container Management - Cluster where the source instance is located - Stateful Workloads : Select the workload redis-a and create a NodePort service with the container port and service port set to 6379.

  2. Verify the service and make sure that the workload selector includes the following labels.

    app.kubernetes.io/component: redis
    app.kubernetes.io/managed-by: redis-operator
    app.kubernetes.io/name: redis-a
    app.kubernetes.io/part-of: redis-failover
    redis-failover: master
    redisfailovers.databases.spotahome.com/name: redis-a
    

Creating Configuration

In Container Management - Cluster where the target instance is located - Configuration and Storage - Configurations , create a configuration item redis-sync for the Redis-shake instance. Import the file sync.toml (file content can be found in the Appendix ), and make sure to modify the following:

  • source.address: The service address of the source instance redis-a created in the previous step:

    address = "10.233.109.145:31278"
    
  • Access password for the source instance: This can be obtained from the Overview page of the Redis instance:

    password = "3wPxzWffdn" # keep empty if no authentication is required
    
  • Address for accessing the target instance. This address can either be the newly created ClusterIP service address or the default Headless service address rfr-redis-b of the instance redis-b . In this example, we use the Headless service address:

    address = "rfr-redis-b:6379"
    

    You can find this service information in Cluster Management - Cluster where the target instance is located - Workloads - Access Method . Similar to the following screenshot:

  • Access password for the target instance. This can be obtained from the Redis instance's Overview page under the Data Services module:

    password = "3wPxzWffdn" # keep empty if no authentication is required
    

    Similar to the following location:

  • Set the target type to standalone :

    [target]
    type = "standalone" # "standalone" or "cluster"
    

Creating Redis-shake

  1. Open the App Workbench , select Wizard - Based on Container Image and create an application Redis-shake-sync :

  2. Fill in the application configuration as per the instructions below.

    • The cluster and namespace of the application should match the Redis instance.
    • Image address:

      release.daocloud.io/ndx-product/redis-shake@sha256:46652d7d8893fa4508c3c6725afc1e211fb9cb894c4dc85e94287395a32fc3dc
      
    • The access type for the default service is set to NodePort, with container port and service port both set to 6379.

    • Advanced Settings - Lifecycle - Startup Command - Enter the run parameter:

      /etc/sync/sync.toml
      
    • Advanced Settings - Data Storage : Add configuration item redis-sync and set the path to:

      /etc/sync
      
    • Advanced Settings - Data Storage : Add a temporary path with the container path set to:

      /data
      
  3. Click OK to complete the creation of Redis-shake.

Once Redis-shake is created, data synchronization between Redis instances will begin. You can verify the synchronization using the redis-cli tool, but we won't go into detail here.

Data Recovery

Diagram: Data Recovery instance redis-b >> instance redis-a

When the source instance instance redis-a comes back online, we need to recover the incremental data from the target instance instance redis-b. This requires deploying another Redis-shake instance on instance redis-a to achieve data recovery from instance redis-b to instance redis-a. The configuration process is similar to the data synchronization process, but in the opposite direction. Once the Redis-shake is created, data recovery will automatically begin.

Note

Before bringing the source instance online, make sure to stop the Redis-shake-sync used for normal synchronization to avoid incorrect synchronization overwriting newly added data.

Restoring the Master-Slave Relationship

To restore the initial master-slave synchronization relationship instance redis-a >> instance redis-b, stop the running Redis-shake-recovery instance in the cluster where instance redis-a is located in the Container Management . Then, restart the Redis-shake-sync instance in the target cluster, and the initial master-slave relationship will be reestablished.

Appendix

sync.toml
type = "sync"

[source]
version = 6.0 # redis version, such as 2.8, 4.0, 5.0, 6.0, 6.2, 7.0, ...
address = "10.233.109.145:6379"
username = "" # keep empty if not using ACL
password = "3wPxzWffdn" # keep empty if no authentication is required
tls = false
elasticache_psync = "" # using when source is ElastiCache. ref: https://github.com/alibaba/RedisShake/issues/373

[target]
type = "standalone" # "standalone" or "cluster"
version = 6.0 # redis version, such as 2.8, 4.0, 5.0, 6.0, 6.2, 7.0, ...
# When the target is a cluster, write the address of one of the nodes.
# redis-shake will obtain other nodes through the __cluster nodes__ command.
address = "10.233.103.2:6379"
username = "" # keep empty if not using ACL
password = "Aa123456" # keep empty if no authentication is required
tls = false

[advanced]
dir = "data"

# runtime.GOMAXPROCS, 0 means use runtime.NumCPU() cpu cores
ncpu = 4

# pprof port, 0 means disable
pprof_port = 0

# metric port, 0 means disable
metrics_port = 0

# log
log_file = "redis-shake.log"
log_level = "info" # debug, info or warn
log_interval = 5 # in seconds

# redis-shake gets key and value from rdb file, and uses RESTORE command to
# create the key in target redis. Redis RESTORE will return a "Target key name
# is busy" error when key already exists. You can use this configuration item
# to change the default behavior of restore:
# panic:   redis-shake will stop when meet "Target key name is busy" error.
# rewrite: redis-shake will replace the key with new value.
# ignore:  redis-shake will skip restore the key when meet "Target key name is busy" error.
rdb_restore_command_behavior = "rewrite" # panic, rewrite or skip

# pipeline
pipeline_count_limit = 1024

# Client query buffers accumulate new commands. They are limited to a fixed
# amount by default. This amount is normally 1gb.
target_redis_client_max_querybuf_len = 1024_000_000

# In the Redis protocol, bulk requests, that are, elements representing single
# strings, are normally limited to 512 mb.
target_redis_proto_max_bulk_len = 512_000_000

Comments