Distributed Cache Service. User Guide. Date

Transcription

1 Date

2 Contents Contents 1 Introduction DCS Application Scenarios Advantages Compatibility with Redis Commands DCS Instance Specifications Accessing DCS How to Access DCS Related Services User Permissions Related Concepts Getting Started Setting Up Environments Creating a DCS Instance Accessing a DCS Instance Viewing Details of a DCS Instance Operation Guide Configuring Redis-Specific Parameters Stopping DCS Instances Starting DCS Instances Restarting DCS Instances Deleting DCS Instances Scaling Up DCS Instances Changing Instance Passwords Modifying an Instance's Maintenance Time Window Modifying an Instance's Security Group Backing Up and Restoring DCS Instances Overview Configuring a Backup Policy Manually Backing Up a DCS Instance Restoring a DCS Instance FAQs ii

3 Contents 4.1 Does DCS Support Common Redis Clients Such as Jedis? Should I Install Redis If I Already Have a DCS Instance? Do I Need to Enter a Password When Accessing a DCS Instance? Where Can I Obtain the Password? Does Standby Cache Node Remain Synchronized With Master Cache Node? Does DCS Support Data Persistence? Can DCS Be Accessed over Public Networks? Can I Modify Configuration Parameters of DCS Instances? Can I Recover Data from Deleted DCS Instances? What Is the Default Data Eviction Policy? What Should I Do If the Connection Between My Redis Client and the DCS Instance Times Out? Why Is Available Memory of Unused DCS Instances Less Than Total Memory and Why Is Memory Usage of Unused DCS Instances Greater Than Zero? Does DCS Support Data Read/Write in Multiple Databases? Why Can't I Back Up or Scale Up My DCS Instances? Do DCS Instances in Cluster Mode Support Native Redis Clusters? Why Do My DCS Instances Sometimes Become Unavailable While They Are in Running State? Why Is the Used Instance Memory Displayed on Cloud Eye Console Slightly Higher than the Available Instance Memory? How to Select and Configure a Security Group? Do DCS Instances Support Cross-VPC Access?...46 A Glossary...48 B Change History iii

4 1 Introduction 1 Introduction 1.1 DCS (DCS) is an online, distributed, in-memory cache service. It is reliable, scalable, usable out of the box, and easy to manage. DCS makes it easy to deploy, operate, and scale in-memory data caches in the cloud. DCS uses Redis 3.0.7, a storage system that supports multiple types of data structures, including key-value pairs. It can be used in scenarios such as caching, event pub/sub, and high-speed queuing. Redis is networked and written in ANSI C. It supports direct read/write of strings, hashes, lists, queues, and sets. Redis works with in-memory datasets and provides persistence. For more information about Redis, visit Clients can access DCS instances only when they are installed on Elastic Cloud Servers (ECSs) and belong to the same virtual private cloud (VPC) as the DCS instances. Figure 1-1 Accessing DCS instances NOTE Like clients, DCS instances are also run on ECSs. A client must belong to the same VPC and subnet as the DCS instance it is attempting to access

5 1 Introduction There are three deployment modes for DCS instances: Single-node Each DCS instance runs on only one cache node. Such DCS instances are mainly used for temporary data storage but are not suitable for high-reliability service scenarios. Single-node DCS instances support intensive concurrent read/write operations, but do not implement data persistence. Data will be deleted after single-node DCS instances are restarted or stopped. Master/standby DCS instances run in master/standby mode, providing higher availability than singlenode DCS instances: Each DCS instance runs on two cache nodes (one master and one standby), and supports data persistence. The standby node is invisible to the client and does not provide data read/write services. It is kept synchronized with the master node through data synchronization. The standby node takes over automatically if the master node fails. Cluster DCS instances run in cluster mode, providing higher instance specifications and performance than master/standby DCS instances: Each DCS instance has at least three master nodes. Compared with master/standby DCS instances, DCS instances in cluster mode support higher instance specifications and concurrency. Each master node has a redundant node. If a master node is faulty, the redundant node becomes the master node and provides data read/write services for users. Each redundant node is invisible to Redis clients, and it is kept synchronized with its master node through data synchronization. 1.2 Application Scenarios For e-commerce, video sharing, gaming, and other data-intensive applications, it is essential to retrieve data fast. The primary purpose of DCS is to provide fast yet inexpensive data retrieval. With DCS, you can retrieve data from in-memory data stores instead of relying entirely on slower disk-based databases. DCS instances are also fully managed. You no longer need to perform management tasks such as hardware provisioning, software patching, monitoring, or failure recovery. Consider using DCS when you need to cache the following types of data: Web pages Caching the content of web pages improves page load times. The cached content can include static data such as Hypertext Markup Language (HTML) pages, Cascading Style Sheets (CSS), and images. Status DCS provides quick and simple queries of session status and application-scaling status, enabling easy sharing of status information. Application objects DCS serves as a level-2 cache at the service layer and provides data storage and access for external entities. Storing frequently requested data in DCS instances can off load databases and reduce access latency

6 1 Introduction Events 1.3 Advantages DCS provides continuous event querying, allowing streams of events to be processed as they occur. The advantages of DCS include: Usability off the shelf You can create DCS instances with just a few clicks on the DCS console or by calling application programming interfaces (APIs). DCS responds to instance creation commands without requiring you to deal with complex underlying software and hardware. Customizable specifications DCS instance specifications range from 1 GB to 512 GB, increasing the flexibility of DCS. Reliability Data persistence, support for master/standby and cluster modes, and automatic failover between cache nodes ensure data reliability and service availability. Ease of maintenance You can create, start, stop, restart, and delete DCS instances with just a few clicks on the DCS console. DCS continuously monitors and maintains DCS instances so that you can focus on application development. Service monitoring metrics include keys, the number of client connections, and processed commands. System monitoring metrics include memory, CPU, and network usage. 1.4 Compatibility with Redis Commands DCS is built on the Redis engine and supports the majority of Redis commands. Any clients that support Redis can access DCS. For security purposes, certain Redis commands are disabled in DCS. For more information about Redis commands, visit commands. NOTE DCS is compatible with Redis 3.0 and supports the GEO command of Redis 3.2. However, DCS does not support certain commands and scripts added in Redis 3.2 and 4.0, which are listed in Commands and Scripts That Are Added in Later Redis Versions but Not Supported by DCS. Some Redis commands, such as FLUSHALL, take a long time to run. While running these commands, DCS instances may not respond to other commands and change to the faulty state. However, after the command finishes executing, the instance will return to normal. The time limit for executing a Redis command is 1 minute. After the execution of a command times out, your client will be automatically disconnected from the server

7 1 Introduction DCS Instances in Single-Node or Master/Standby Mode Table 1-1 Redis commands disabled in single-node and master/standby DCS instances Category Keys Server Commands MIGRATE SLOWLOG, SLAVEOF, SHUTDOWN, SAVE, MONITOR, LASTSAVE, DEBUG SEGFAULT, DEBUG OBJECT, CONFIG SET, CONFIG REWRITE, CONFIG RESETSTAT, COMMAND INFO, COMMAND GETKEYS, COMMAND COUNT, COMMAND, BGSAVE, BGREWRITEAOF Redis clients receive "(error) ERR unknown command" from DCS and probably exhibit unpredictable behavior after running the disabled commands. DCS cannot control their behavior. For example, if the MONITOR command is run, redis-cli will keep waiting for monitoring results even after receiving "(error) ERR unknown command" from DCS and refuse to run other commands during the wait period. DCS Instances in Cluster Mode DCS instances in cluster mode do not support the following commands: Redis commands disabled for single-node and master/standby DCS instances Commands listed in Table 1-2 Table 1-2 Redis commands disabled in cluster DCS instances Category Server Script Transaction Connection Keys Cluster codis Commands SYNC, PSYNC SCRIPT DEBUG DISCARD, EXEC, MULTI, UNWATCH, WATCH SELECT NOTE Parameters of the SELECT command can only be set to 0. MOVE, RESTORE CLUSTER PSUBSCRIBE(), PUBLISH, PUNSUBSCRIBE, SUBSCRIBE, RANDOMKEY, UNSUBSCRIBE, TIME, SLOTSINFO, SLOTSDEL, SLOTSMGRTSLOT, SLOTSMGRTONE, SLOTSCHECK, SLOTSMGRTTAGSLOT, SLOTSMGRTTAGONE

8 1 Introduction Table 1-3 lists the commands supported by cluster DCS instances for multi-key operations in the same slot. Table 1-3 Commands supported by cluster DCS instances for multi-key operations in the same slot Category Set SortedSet HyperLogLog List Keys String Commands SINTERSTORE, SINTER, SUNIONSTORE, SUNION, SDIFFSTORE, SDIFF, SMOVE ZUNIONSTORE, ZINTERSTORE PFCOUNT, PFMERGE BLPOP, BRPOP, BRPOPLPUSH RENAME, RENAMENX, BITOP, RPOPLPUSH MSETNX

9 1 Introduction Redis clients receive "(error) ERR unknown command" from DCS and probably exhibit unpredictable behavior after running the disabled commands. DCS cannot control their behavior. For example, if the PSUBSCRIBE, SUBSCRIBE, or MONITOR command is run, redis-cli will keep waiting for responses from DCS even after receiving "(error) ERR unknown command" from DCS and refuse to run other commands during the wait period. When the FLUSHDB or FLUSHALL command is run, execution of other service commands may be blocked for a long time in case of a large amount of cached data. When the KEYS command is run, execution of other service commands may be blocked for a long time or even extra high memory may be occupied in case of a large amount of cached data. Therefore, when running the KEYS command, describe the exact pattern and do not use fuzzy keys *. keys * can be used during commissioning or when the number of keys does not exceed 5 million. Otherwise, the service cannot run properly. The hash tag mechanism can be used to ensure that multiple keys belong to the same slot. For details, visit When the EVAL or EVALSHA command is run, a DCS cluster uses the first key to compute slots. Ensure that the keys to be operated in your code are in the same slot. For details, visit For the EVAL command: You are advised to learn the Lua script feature of Redis before running the EVAL command. The execution timeout time of a Lua script is 5 seconds. Time-consuming statements such as long-time sleep and large loop statements should be avoided. When calling a Lua script, do not use random functions to specify keys. Otherwise, the execution results are inconsistent on the master and standby nodes. Cluster instances purchased before July 10, 2018 must be upgraded to support the following commands: SINTER, SDIFF, SUNION, PFCOUNT, PFMERGE, SINTERSTORE, SUNIONSTORE, SDIFFSTORE, SMOVE, BLPOP, BRPOP, BRPOPLPUSH, ZUNIONSTORE, ZINTERSTORE, EVAL, EVALSHA, BITOP, RENAME, RENAMENX, RPOPLPUSH, MSETNX, SCRIPT LOAD, SCRIPT KILL, SCRIPT EXISTS, SCRIPT FLUSH. Commands and Scripts That Are Added in Later Redis Versions but Not Supported by DCS Table 1-4 Commands and scripts that are added in later Redis versions but not supported by DCS Command/Script SPOP HSTRLEN Description Command used to control the number of randomly deleted elements. This command is added in Redis 3.2. Parameter not supported: count. Command used to return the value length of hash data. This command is added in Redis

10 1 Introduction Command/Script BITFIELD SWAPDB MEMORY Description Command used to calculate character strings as arrays of binary digits. This command is added in Redis 3.2. Command used to exchange the two databases specified in a DCS Redis instance. This command is added in Redis 4.0. Command used to query and manage the memory usage, for example, query the memory usage of a key, query the details of overall memory usage, apply for memory release, query the internal status of a memory allocator. This command is added in Redis DCS Instance Specifications For each single-node DCS instance, the available memory is less than the total memory because some memory is reserved for system overhead. For each master/standby DCS instance, the available memory is less than the total memory because some memory is reserved for data persistence. If DCS instances are deployed in master/standby mode, only the master cache node contributes to the available memory. Table 1-5 Specifications of single-node and master/standby DCS instances Memory (GB) Total Available (Single Node) Available (Master/ Standby) Maximum Number of Connections Allowed Maximum Intranet Bandwidth (Mbit/s) , , , , , ,

11 1 Introduction Table 1-6 Specifications of cluster DCS instances Total Memory (GB) Available Memory (GB) Maximum Number of Connections Allowed Maximum Intranet Bandwidth (Mbit/s) , , , , NOTE The maximum intranet bandwidth is the sum of maximum inbound and outbound intranet bandwidths. Usually, DCS instances are read more frequently than they are written. Therefore, it is recommended that the maximum inbound intranet bandwidth be one third of the maximum intranet bandwidth and the maximum outbound intranet bandwidth be two thirds of the maximum intranet bandwidth. 1.6 Accessing DCS How to Access DCS There are two ways to access DCS: Using the management console The public cloud provides a web-based management console. To launch the DCS console, create an account on the management console and log in, and then choose Database > on the homepage. On the DCS console, you can create, start, stop, restart, and delete DCS instances, and change their passwords. Using APIs Related Services Virtual Private Cloud (VPC) If you want to integrate DCS into a third-party system for secondary development, use APIs to access DCS. For more information about available APIs, see the Distributed Cache Service API Reference. The VPC service enables users to create private, isolated virtual networks. DCS instances run in VPCs and use the IP addresses and bandwidths of VPCs. VPCs are based on security groups. A security group is a set of access control rules that implements access control for mutually trusted ECSs with the same security protection requirements in the same VPC

12 1 Introduction Elastic Cloud Server (ECS) Cloud Trace Service (CTS) The ECS service provides scalable, on-demand cloud servers for secure, flexible, and efficient application environments, ensuring service reliability. After you create DCS instances, you can connect to them through ECSs. CTS provides a history of operations performed on cloud service resources. With CTS, you can query, audit, and review operations. Traces include operation time, resource objects, resource IDs, requesters' IP addresses, resource operation requests, and responses. Currently, CTS records the following operations on DCS instances: Creating, starting, stopping, restarting, and deleting DCS instances Configuring Redis-specific parameters Changing instance passwords Modifying basic information Identity and Access Management (IAM) Cloud Eye IAM provides identity authentication and permission management. It is used to authenticate access to DCS. Cloud Eye is a secure, scalable monitoring platform in the public cloud. It monitors DCS service metrics and sends notifications if alarms or events occur. Table 1-7 DCS monitoring metrics displayed on the Cloud Eye console Metric Description Value Range Monitored Object CPU Usage Memory Usage CPU consumed by monitored objects Memory consumed by monitored objects 0% to 100% ECS 0% to 100% ECS Network Input Throughput Inbound throughput per second on a port 0 bytes/s ECS Network Output Throughput Outbound throughput per second on a port 0 bytes/s ECS Connected Clients Number of connected clients, including connections established by the DCS server (a management service in the background) to monitor DCS instances and excluding connections from standby nodes 0 DCS

13 1 Introduction Metric Description Value Range Monitored Object Client Longest Output List Longest output list among current client connections 0 DCS Client Biggest Input Buf Maximum input data length among current client connections 0 bytes/s DCS Blocked Clients Number of clients suspended by block operations, such as BLPOP, BRPOP, and BRPOPLPUSH 0 DCS Used Memory Total number of bytes allocated by Redis using its allocator 0 bytes/s DCS Used Memory RSS Number of bytes that Redis allocates as seen by the operating system (a.k.a resident set size). It includes all stack and heap memory but not swapped-out memory. 0 bytes/s DCS Used Memory Peak Peak memory consumed by Redis since the Redis server last started 0 bytes/s DCS Used Memory Lua Number of bytes used by the Lua engine 0 bytes/s DCS

14 1 Introduction Metric Description Value Range Monitored Object Memory Fragmentation Ratio Ratio between Used Memory RSS and Used Memory. Ideally, the Used Memory RSS should be only slightly higher than Used Memory. When Used Memory RSS is much greater than Used Memory, there is memory fragmentation (internal or external), which can be evaluated by checking Memory Fragmentation Ratio. When Used Memory is much greater than Used Memory RSS, part of Redis memory has been swapped out by the operating system. In this case, significant latency can be expected. Because Redis does not have control over how its allocations are mapped to memory pages, high Used Memory RSS is often the result of a spike in memory usage. 0 DCS Total Connections Received Total number of connections the Redis server has received since it last started, including connections established by the DCS server (a management service in the background) to monitor DCS instances 0 DCS Total Commands Processed Total number of commands the Redis server has processed since it last started, including commands issued by the DCS server (a management service in the background) to monitor DCS instances 0 DCS Instantaneous Ops per Second Number of commands processed per second 0 DCS Total Net Input Bytes Total number of bytes the Redis server has received since it last started 0 bytes/s DCS

15 1 Introduction Metric Description Value Range Monitored Object Total Net Output Bytes Total number of bytes the Redis server has sent since it last started 0 bytes/s DCS Instantaneous Input Kbps Instantaneous Output Kbps Instantaneous input traffic 0 kbit/s DCS Instantaneous output traffic 0 kbit/s DCS Rejected Connections Number of connections that have exceeded maxclients and been rejected by the Redis server since it last started 0 DCS Sync Full Total number of full synchronizations since the Redis server last started 0 DCS Sync Partial OK Total number of successful incremental synchronizations since the Redis server last started 0 DCS Sync Partial Err Total number of failed incremental synchronizations since the Redis server last started 0 DCS Expired Keys Total number of keys that have expired since the Redis server last started 0 DCS Evicted Keys Number of keys that have been evicted due to insufficient memory since the Redis server last started 0 DCS Keyspace Hits Total number of query hits in the master dictionary since the Redis server last started 0 DCS Keyspace Misses Total number of query misses in the master dictionary since the Redis server last started 0 DCS PubSub Channels Number of Pub/Sub channels 0 DCS PubSub Patterns Number of Pub/Sub patterns 0 DCS Aof Current Size AOF file size 0 DCS

16 1 Introduction Metric Description Value Range Monitored Object Latest Fork usec Duration of the latest fork operation 0 ms DCS User Permissions The public cloud system provides two types of user permissions by default: user management and resource management. User management refers to the management of users, user groups, and user group rights. Resource management refers to the control operations that can be performed by users on cloud service resources. For details, see Permissions. 1.7 Related Concepts Region Availability Zone Project A region is a geographic area in which DCS is located. DCS services in the same region can communicate with each other over an intranet, but those in different regions cannot. Public cloud data centers are deployed in different places around the world. DCS is therefore available in different regions. For example, applications can be designed to meet requirements of users in specific regions or comply with local laws or regulations. Each region contains many availability zones (AZs), in which power and networks are physically isolated. AZs in the same region can communicate with each other over an intranet. Each AZ provides cost-effective and low-latency network connections that are unaffected by faults that may occur in other AZs. Using DCS deployed in an independent AZ protects your applications against failures that occur in a single place. Projects are used to group and isolate OpenStack resources (computing, storage, and network resources). A project can be a department or a project team. Multiple projects can be created for one account

17 2 Getting Started 2 Getting Started 2.1 Setting Up Environments Procedure To access DCS instances through a VPC, create a VPC and configure security groups and subnets for it before using DCS. A VPC provides an isolated, user-configurable, and user-manageable virtual network environment for DCS. Using VPCs enhances public cloud resource security and simplifies network deployment. Once you have created a VPC, you can use it for all DCS instances you subsequently create. Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Click Service List, and choose Network > Virtual Private Cloud to launch the VPC console. On the Dashboard page, click Create VPC. Create a VPC as guided by the VPC console, retaining the default values of the VPC parameters unless otherwise required. For more information on how to create a VPC, see the Virtual Private Cloud. After a VPC is created, a subnet is also created in the subnet. If the VPC needs more subnets, go to Step 6. Otherwise, go to 7. Step 6 In the navigation pane, choose Virtual Private Cloud. On the Virtual Private Cloud page, click the name of the VPC in which you intend to create a subnet. On the Subnets tab page, click Create Subnet. Create a subnet as guided by the VPC console, retaining the default values of the subnet parameters unless otherwise required. For more information on how to create a subnet, see the Virtual Private Cloud

18 2 Getting Started Step 7 In the navigation pane, choose Security Group. On the Security Group page, click Create Security Group. Create a security group for the VPC, retaining the default values of the security group parameters unless otherwise required. For more information on how to create a security group, see the Virtual Private Cloud User Guide. ----End 2.2 Creating a DCS Instance Scenario DCS can be used only after DCS instances are created. DCS supports three types of DCS instances: single-node, master/standby, and cluster. DCS does not provide built-in encryption. Encrypt any sensitive data before transmitting or storing it. The purpose, scope, processing method, and time limits of data processed by DCS must comply with local laws and regulations. Prerequisites The VPC where the DCS instance is to be created is available. Security groups and subnets have been configured for the VPC. For more information on how to create VPCs, security groups, and subnets, see Setting Up Environments or the Virtual Private Cloud. Procedure Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Step 6 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. On the Cache Manager page, click Create DCS Instance. Specify DCS instance parameters, such as the name, AZ, VPC, security group, and subnet

19 2 Getting Started Table 2-1 DCS instance parameters Parameter Region Name Description AZ Cache Engine Engine Version Instance Type Instance Specification VPC Subnet Description Region in which DCS is located. To select a different region, use the region selector at the upper left of the main menu bar. Name of the new DCS instance. A DCS instance name cannot be left unspecified and must: Consist of 4 to 64 characters. Consist of only letters, digits, hyphens (-), and underscores (_). Start with an uppercase or lowercase letter. Description of the new DCS instance. AZ in which the new DCS instance resides. Cache engine used. Currently, only Redis is supported. Cache engine version. Currently, only Redis is supported. Type of the DCS instance being created. Currently, three types are supported: single-node, master/standby, and cluster. Select an instance type based on your service scenario. For more information, see DCS Instance Type. NOTE Cluster DCS instances do not support cross-vpc access through VPC peering connections. Total memory of the new DCS instance. Value: 64, 128, 256, or 512 GB if the DCS instance is in cluster mode 1, 2, 4, 8, 16, 32, or 64 GB if the DCS instance is in single-node or master/standby mode By default, each user can create a maximum of five DCS instances and use a total of 400 GB memory. When creating a DCS instance on the DCS console, available free memory is listed below the Instance Specification field. If you need more memory, click Increase quota below the Instance Specification field and contact customer service. For more information, see How Can I Apply for a Higher Quota?. VPC in which the new DCS instance resides. A VPC provides an isolated, user-configurable, and user-manageable virtual network environment for your DCS instances. Click View VPC to show more details of the chosen VPC, including security group rules. Name and IP address range of the subnet in which the new DCS instance resides

20 2 Getting Started Parameter Security Group Password Time Window Backup Policy Description Security group that controls access to the new DCS instance. A security group is a set of access control rules that implements access control for mutually trusted ECSs with the same security protection requirements in the same VPC. Password required for accessing the new DCS instance. NOTE For security purposes, the system prompts you to enter an instance-specific password when you are accessing the DCS instance. Keep your instance password secure and change it periodically. Passwords cannot be left unspecified and must: Consist of 8 to 32 characters. Contain at least three of the following character types: uppercase letters, lowercase letters, digits, and special characters (`~!@#$ %^&*()-_=+\ [{}]:'",<.>/?). Time range for any scheduled maintenance activities to occur for cache nodes of this DCS instance. Values: 22:00-02:00 02:00-06:00 06:00-10:00 10:00-14:00 14:00-18:00 18:00-22:00 Data backup policy. This parameter is displayed only when the instance type is master/standby. For more information on how to configure a backup policy, see Configuring a Backup Policy. Step 7 Click Create Now. The Confirm tab page is displayed, providing information about the new DCS instance, including the instance name, cache engine version, and instance specifications. Step 8 Step 9 On the Confirm tab page, review the instance information and click Submit. After the new DCS instance has been created, return to the Cache Manager page to view and manage your DCS instances. It takes 5 to 15 minutes to create a DCS instance. However, it will take up to approximately 30 minutes if the DCS instance is in cluster mode. After a DCS instance has been successfully created, it enters the Running state by default

21 2 Getting Started NOTE ----End If the new DCS instance failed to be created, delete the unsuccessful instance creation task by following the procedure in Deleting Instance Creation Tasks That Have Failed to Run. Then, create the DCS instance again. If the DCS instance still fails to be created, contact customer service. 2.3 Accessing a DCS Instance You can access the created DCS instances through your Redis client. DCS works with multiple types of Redis clients. This section describes how to use the rediscli and Redis Java (Jedis) clients to access a DCS Redis instance. For more information on how to use other Redis clients, visit Prerequisites The DCS Redis instance you want to access is in the Running state. An ECS has been created to serve as your Redis client. For more information on how to create ECSs, see the Elastic Cloud Server. The GNU Compiler Collection (GCC) has been installed on the ECS serving as your Redis client. NOTE The ECS serving as your Redis client and the DCS instance to be accessed must belong to the same VPC and can communicate with each other. Procedure Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Step 6 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. Obtain the IP address:port number of the DCS instance. 1. On the Cache Manager page, click the name of the DCS instance you want to access. A page with details of the DCS instance is displayed. 2. On the Summary tab page of the instance details page, view the Connection Address (IP address:port number) of the DCS instance. Access the chosen DCS instance. By using redis-cli a. Download the source code package of your Redis client from download.redis.io/releases/redis tar.gz. b. Upload the source code package of your Redis client to the ECS serving as your Redis client

22 2 Getting Started c. Log in to the ECS that serves as your Redis client. d. Run the following command to decompress the redis directory from the source code package of your Redis client: tar -xzf redis tar.gz e. Run the following commands to go to the redis directory and compile the source code of your Redis client: cd redis make f. Run the following commands to access the chosen DCS instance: cd src./redis-cli -h p 6379./redis-cli -h ip-address -p port-number In the second command, use the DCS instance IP address and port number obtained in Step 5. g. On redis-cli, run the following command to authenticate access to the selected DCS instance: auth password In this command, password indicates the password used for login to the chosen DCS instance. This password is defined during DCS Redis instance creation. If information similar to the following is displayed, access to the chosen DCS instance is authenticated, and you can now read from and write to the chosen DCS instance: ~]#./redis-cli -h p :6379> auth ****** OK :6379> By using Jedis In this example command output: is an example IP address of DCS instance, which is obtained in Step is an example port number of DCS instance, which is obtained in Step 5. a. Obtain the source code of the Jedis client from b. Write code. i. Example code for a single Jedis connection //Create a connection String host = " "; int port = 6379; String pwd = "passwd"; ii. Jedis client = new Jedis(host, port); client.auth(pwd); client.connect(); //Run the set command String result = client.set("key-string", "Hello, Redis!"); System.out.println( String.format("set command result:%s", result) ); //Run the get command String value = client.get("key-string"); System.out.println( String.format("get command result:%s", value) ); Example code for a Jedis connection pool //Generate configuration information of a Jedis connection pool String ip = " "; int port = 6379;

23 2 Getting Started ----End String pwd = "passwd"; GenericObjectPoolConfig config = new GenericObjectPoolConfig(); config.settestonborrow(false); config.settestonreturn(false); config.setmaxtotal(100); config.setmaxidle(100); config.setmaxwaitmillis(2000); JedisPool pool = new JedisPool(config, ip, port, , pwd);// Generate a Jedis connection pool when the client application is being initialized //Get a Jedis connection from the Jedis connection pool when the client initiates a request Jedis client = pool.getresource(); try { //Run commands String result = client.set("key-string", "Hello, Redis!"); System.out.println( String.format("set command result:%s", result) ); String value = client.get("key-string"); System.out.println( String.format("get command result:%s", value) ); } catch (Exception e) { // TODO: handle exception } finally { //Return the Jedis connection to the Jedis connection pool after the client's request is processed if (null!= client) { pool.returnresource(client); } } // end of try block //Destroy the Jedis pool when the client application is closed pool.destroy(); c. Compile code according to the readme file in the source code of the Jedis client. Run the Jedis client to access the chosen DCS Redis instance. 2.4 Viewing Details of a DCS Instance Scenario On the DCS console, you can view details about a DCS instance. Procedure Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. On the Cache Manager page, filter DCS instances by instance status and/or name to find the desired DCS instance

24 2 Getting Started Table 2-2 DCS instance statuses State Creating Running Faulty Starting Restarting Stopping Stopped Scaling up Description The instance is being created. The instance is running properly. Only instances in the Running state can provide in-memory cache service. The instance is not running properly. The instance is being started. The instance is being restarted. The instance is being stopped. The instance has been stopped gracefully. Instances in the Stopped state no longer provide in-memory cache service. The instance is being scaled up. Step 6 On the DCS instance list, click the name of the chosen DCS instance to display more details about it. Table 2-3 Parameters on the DCS instance details page Category Parameter Description Basic Informatio n Name Instance Type Instance Specification (GB) Used/Available Memory (MB) Status Time Window Name of the chosen instance. To modify the instance name, click the icon. Type of the chosen instance. Currently, three types are supported: single-node, master/standby, and cluster. Specification of the chosen instance. The used memory space and maximum allowed memory space of the chosen instance. The used memory space includes: Size of data stored on the DCS instance Size of Redis-server buffers (including client buffer and repli-backlog) and internal data structures Status of the chosen DCS instance. For details, see Table 2-2. Time range for any scheduled maintenance activities to occur for cache nodes of this DCS instance. To modify the maintenance time window, click the Time Window. icon next to

25 2 Getting Started Category Parameter Description ID Connection Address Created Description Monitoring metrics ID of the chosen instance. IP address and port number of the chosen instance. Time at which the chosen instance was created. Description of the selected instance. To modify the description, click the icon. Monitoring metrics of the chosen instance. Click the View icon to switch to the Cloud Eye console on which monitoring metrics of the selected DCS instance are displayed. Network AZ Availability zone in which the cache node running the chosen instance resides. Security Group VPC Subnet Security group that controls access to the chosen instance. To modify the security group, click the next to Security Group. VPC in which the chosen instance resides. Subnet in which the chosen instance resides. icon ----End

26 3 Operation Guide 3 Operation Guide This chapter describes how to manage DCS instances. In the event that a cache node of a DCS instance is faulty: The instance remains in the Running state and you can continue to read from and write to the instance. This is achieved thanks to the high availability of DCS. Cache nodes can recover from internal faults automatically. Manual fault recovery is also supported. Certain operations (such as backup, restoration, parameter configuration, and scaling) in the management zone are not supported during fault recovery. You can contact customer service or perform these operations after cache nodes recover from faults. 3.1 Configuring Redis-Specific Parameters Prerequisites Procedure On the DCS console, you can configure Redis-specific parameters to keep DCS instances performing optimally. NOTE The default values are already optimized for typical use cases. Reconfiguring parameters is recommended only when necessary. The DCS instance you want to configure is in the Running state, and the instance type is single-node or master/standby (Redis parameters of DCS instances in cluster mode are not modifiable). Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project

27 3 Operation Guide Step 3 Click Service List, and choose Database > to launch the DCS console. Step 4 Step 5 In the navigation pane, choose Cache Manager. On the Cache Manager page, click the name of the DCS instance you want to configure. A page with details of the DCS instance is displayed. Step 6 Step 7 Step 8 On the instance details page, click the Configuration Parameters tab. On the Configuration Parameters tab page, click Modify. Modify Redis-specific parameters based on your requirements. Table 3-1 Redis-specific parameters Parame ter Description Possible Values Default Value maxme morypolicy How DCS will select what to remove when maxmemory is reached. For more information about this parameter, see What Is the Default Data Eviction Policy? volatile-lru allkeys-lru volatile-random allkeys-random volatile-ttl noeviction noeviction hashmaxziplistentries hashmaxziplistvalue listmaxziplistentries listmaxziplistvalue set-maxintsetentries When the number of entries in hashes is less than the value of this parameter, hashes are encoded using a memory efficient data structure. When the biggest entry in hashes does not exceed the length threshold indicated by this parameter, hashes are encoded using a memory efficient data structure. When the number of entries in lists is less than the value of this parameter, lists are encoded using a memory efficient data structure. When the biggest entry in lists does not exceed the length threshold indicated by this parameter, lists are encoded using a memory efficient data structure. When a set is composed entirely of strings that happen to be integers in radix 10 in the range of 64 bit signed integers, sets are encoded using a memory efficient data structure. 1 to to to to to

28 3 Operation Guide Parame ter Description Possible Values Default Value zsetmaxziplistentries zsetmaxziplistvalue latencymonitorthreshol d When the number of entries in sorted sets is less than the value of this parameter, sorted sets are encoded using a memory efficient data structure. When the biggest entry in sorted sets does not exceed the length threshold indicated by this parameter, sorted sets are encoded using a memory efficient data structure. Only events that run in more time than the configured latency-monitorthreshold will be logged as latency spikes. If the latency-monitor-threshold is set to 0, latency monitoring is disabled. If the latency-monitor-threshold is set to a value greater than 0, all events blocking the server for a time equal to or greater than the configured latency-monitorthreshold will be logged. By running the LATENCY command, you can perform operations related to latency monitoring, such as enabling latency monitoring, reporting the latest latency events logged, and obtaining statistical data. For more information about the latency-monitor-threshold, visit 1 to to to ms 0 reserved - memory The number of megabytes reserved for the backend to perform internal processing such as persistence and master/standby replication. This parameter is configurable only for master/standby instances. 0% to 50% of maximum memory space initially available to the instance and below the current free memory space NOTE: For more information about maximum available memory of each instance type, see Table

29 3 Operation Guide Parame ter Description Possible Values Default Value timeout Connection between the client and server (DCS instance) will be closed if the client is idle for the timeout period (measured in seconds). A timeout period of 0 seconds indicates that the timeout function is disabled. 0 to 7200 seconds

30 3 Operation Guide Parame ter Description Possible Values Default Value notifykeyspac e-events Keyspace event notification. If this parameter is configured, the Redis Sub/Pub feature will allow clients to receive an event when a Redis data set is modified. If the parameter value is an empty character string, keyspace event notification is disabled. If the parameter value is a string of multiple characters, keyspace event notification is enabled and each character identifies a class of keyspace events for which Redis will send notifications. "" NOTE The parameter value must contain either K or E. A is an alias for "g $lshzxe"and cannot be used together with any of the characters "g $lshzxe". For example, the value Kl means that Redis will notify Pub/Sub clients about keyspace events and list commands. The value AKE means Redis will notify Pub/Sub clients about all events. K: Keyspace events, published with the keyspace@ prefix E: Keyevent events, published with keyevent@ prefix g: Generic commands (non-type specific) such as DEL, EXPIRE, and RENAME $: String commands l: List commands s: Set commands h: Hash commands z: Sorted set commands x: Expired events (events generated every time a key expires)

31 3 Operation Guide Parame ter Description Possible Values Default Value e: Evicted events (events generated when a key is evicted for maxmemory) A: Alias for "g$lshzxe", so that the "AKE" string means all the events. Note that the parameter value must contain either K or E. NOTE For more information about Redis-specific parameters, visit The latency-monitor-threshold parameter is usually used for fault location. After locating faults based on the latency information collected, change the value of latency-monitor-threshold to 0 to avoid unnecessary latency. Step 9 Step 10 After you have finished setting parameters, click Save. Click OK to confirm. ----End 3.2 Stopping DCS Instances Scenario On the DCS console, you can stop one or multiple DCS instances at a time. After a DCS instance is stopped, it can no longer be read from or written to. If it is a single-node instance, data will also be deleted from it when it is stopped. An attempt to stop a DCS instance while it is being backed up may result in a failure. Prerequisites Procedure The DCS instances you want to stop are in the Running state. Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project

32 3 Operation Guide Step 3 Click Service List, and choose Database > to launch the DCS console. Step 4 Step 5 Step 6 In the navigation pane, choose Cache Manager. Choose instances to stop. To stop a single instance, click Stop in the same row as the instance. To stop multiple instances, select the instances to stop and click Stop above the instance list. In the Stop dialog box, click OK to confirm that you want to stop the instances. It takes 1 to 30 minutes to stop DCS instances. After DCS instances are stopped, their status changes from Running to Stopped. ----End 3.3 Starting DCS Instances Scenario On the DCS console, you can start one or multiple DCS instances at a time. Prerequisites The DCS instances you want to start are in the Stopped state. Procedure Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Step 6 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. Choose instances to start. To start a single instance, click Start in the same row as the instance. To start multiple instances, select the instances to start and click Start above the instance list. In the Start dialog box, click OK to confirm that you want to start the instances. It takes 1 to 30 minutes to start DCS instances. After DCS instances are started, their status changes from Stopped to Running. ----End

33 3 Operation Guide 3.4 Restarting DCS Instances Scenario On the DCS console, you can restart one or multiple DCS instances at a time. While a DCS instance is restarting, it cannot be read from or written to. If the DCS instance is a single-node instance, data will also be deleted from it when it is restarted. An attempt to restart a DCS instance while it is being backed up may result in a failure. Prerequisites Procedure The DCS instances you want to restart are in the Running or Faulty state. Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Step 6 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. Choose instances to restart. To restart a single instance, choose More > Restart in the same row as the instance. To restart multiple instances, select the instances to restart and click Restart above the instance list. In the Restart dialog box, click OK to confirm that you want to restart the instances. By default, only the instance process will restart. However, selecting Forced Restart will restart the VM on which the chosen DCS instance runs. It takes 1 to 30 minutes to restart DCS instances. After DCS instances are restarted, their status changes to Running. ----End

34 3 Operation Guide 3.5 Deleting DCS Instances Scenario On the DCS console, you can delete one or multiple DCS instances at a time. You can also delete all instance creation tasks that have failed to run. After a DCS instance is deleted, data from the instance will also be deleted without backup. Prerequisites The DCS instances you want to delete have been created. The DCS instances you want to delete are in the Running, Faulty, or Stopped state. Deleting DCS Instances Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Step 6 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. Choose instances to delete. To delete a single instance, choose Operation > More > Delete in the same row as the instance. To delete multiple instances, select the instances to delete and click Delete above the instance list. DCS instances in the Creating, Starting, Stopping, or Restarting state cannot be deleted. In the Delete dialog box, click OK to confirm that you want to delete the instances. It takes 1 to 30 minutes to delete DCS instances. ----End Deleting Instance Creation Tasks That Have Failed to Run Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Click Service List, and choose Database > to launch the DCS console

35 3 Operation Guide Step 4 In the navigation pane, choose Cache Manager. If there are DCS instances that have failed to be created, Instance Creation Failures is displayed above the instance list. Step 5 Click the icon or the number of failed tasks next to Instance Creation Failures. The Instance Creation Failures dialog box is displayed. Step 6 Choose failed instance creation tasks to delete. To delete a single failed task, click Delete in the same row as the task. To delete all failed tasks, click Delete All above the task list. ----End 3.6 Scaling Up DCS Instances Scenario On the DCS console, you can scale up DCS instances to desired specifications. DCS instances in cluster mode do not support scale-up. It takes 5 to 30 minutes to scale up a DCS instance. To minimize the impact of any unexpected results, we recommend scaling up your instance during off-peak periods. During scale-up of single-node DCS instances, you cannot read from or write to DCS, but during scale-up of master/standby DCS instances, you can read from or write to DCS. After successful scale-up, the scaled DCS instances are billed based on new specifications. Prerequisites Procedure The DCS instance you want to scale up is in the Running state. Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Step 6 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. Choose More > Scale up in the same row as the DCS instance you want to scale up. In the Scale Up DCS Instance dialog box, select the specifications to which the chosen DCS instance will scale up. The maximum storage space for both single-node and master/standby instances is 64 GB

36 3 Operation Guide Step 7 In the Scale Up DCS Instance dialog box, click OK. The Apply for Instance Scale-Up page is displayed. This page displays instance information such as the instance name, cache engine version, and new specifications. Step 8 Review instance information. Click Submit on the Confirm tab page. It takes 5 to 30 minutes to scale up a DCS instance. After scale-up is successful, the DCS instance enters the Running state. NOTE ----End If scaling up a single-node DCS instance fails: The instance continues to use the original specifications. In addition, the instance is temporarily unavailable for use and does not support management operations (such as parameter configuration and scale-up). However, after O&M personnel click Retry Scale-Up on the DCS OM console, the instance will be forcibly scaled to the new specifications and become available for use again. If scaling up a master/standby DCS instance fails: Management operations (such as backup, restoration, scale-up, and parameter configuration) are temporarily unsupported. The instance continues to be available for use with its original specifications. Remember not to read or write more data than allowed by the original specifications; otherwise, data loss may occur. After scale-up is successful, the new specifications of the instance take effect. 3.7 Changing Instance Passwords Scenario On the DCS console, you can change the password required for accessing your DCS instance. Prerequisites The DCS instance for which you want to change the password is in the Running state. Procedure Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Step 6 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. Choose More > Change Password in the same row as the chosen instance. In the Change Password dialog box, enter the old password and new password

37 3 Operation Guide NOTE After 5 consecutive incorrect password attempts, the account for accessing the chosen DCS instance will be locked for 5 minutes. Passwords cannot be changed during the lockout period. A DCS instance password cannot be left unspecified and must: Be different from the old password. Consist of 8 to 32 characters. Contain at least three of the following character types: uppercase letters, lowercase letters, digits, and special characters (`~!@#$%^&*()-_=+\ [{}]:'",<.>/?). Step 7 In the Change Password dialog box, click OK to confirm the password change. ----End 3.8 Modifying an Instance's Maintenance Time Window Prerequisite Procedure On the DCS console, after creating a DCS instance, you can modify the maintenance time window of the DCS instance on the instance's Summary page. At least one DCS instance has been created. Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. NOTE Select the same region as the application service of the tenant. Step 3 Step 4 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. A list of DCS instances is displayed on the Cache Manager page. Step 5 Click the name of the DCS instance for which you want to modify the maintenance time window. Step 6 Click the Summary tab. In the Basic Information area, click the icon next to the Time Window parameter. Step 7 Select a new maintenance time window from the drop-down list. Click to save the modification or to cancel. The modification will take effect immediately, that is, the new maintenance time window will appear on the Summary tab page immediately. ----End

38 3 Operation Guide 3.9 Modifying an Instance's Security Group Prerequisites Procedure On the DCS console, after creating a DCS instance, you can modify the security group of the DCS instance on the instance's Summary page. At least one DCS instance has been created. Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. NOTE Select the same region as the application service of the tenant. Step 3 Step 4 Step 5 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. A list of DCS instances is displayed on the Cache Manager page. Click the name of the DCS instance for which you want to modify the security group. Step 6 Click the Summary tab. In the Network area, click the icon next to the Security Group parameter. Step 7 Select a new security group from the drop-down list. Click to save the modification or to cancel the modification. The modification will take effect immediately, that is, the new security group will appear on the Summary tab page immediately. ----End 3.10 Backing Up and Restoring DCS Instances Overview Need for DCS Instance Backup There is a small chance that dirty data could exist in a DCS instance owing to service system exceptions or problems in loading data from persistence files. In addition, some systems demand not only high reliability but also data security, data restoration, and even permanent data storage. Currently, data in DCS instances can be backed up to OBS. If a DCS instance becomes faulty, data in the instance can be restored from backup so that service continuity is not affected

39 3 Operation Guide Backup Modes DCS instances support the following backup modes: Scheduled backup You can create a scheduled backup policy on the DCS console. Then, data in the chosen DCS instances will be automatically backed up at the scheduled time. You can choose the days of the week on which scheduled backup will run. Backup data will be retained for a maximum of seven days. Backup data older than seven days will be automatically deleted. The primary purpose of scheduled backups is to create complete data replicas of DCS instances so that the instance can be quickly restored if necessary. Manual backup Backup requests can also be issued manually. Then, data in the chosen DCS instances will be permanently backed up to OBS. Backup data can be deleted manually. Before performing high-risk operations, such as system maintenance or upgrade, DCS instance data needs to be backed up. Additional Information About Data Backup Instance type Currently, only master/standby DCS instances can be backed up and restored. Singlenode and cluster DCS instances do not support backup or restoration. Working principle Instance data is persisted using the Redis Append Only Files (AOF) feature. Backup tasks are run on standby cache nodes. DCS instance data is backed up by compressing and storing the data persistence files from the standby cache node to OBS. DCS checks instance backup policies once an hour. If a backup policy is matched, DCS runs a backup task for the corresponding DCS instance. Impact on DCS instances during backup DCS instances can continue to provide services during backup. In the event of full-data synchronization or heavy instance load, it takes a few minutes to complete data synchronization. If instance backup starts before data synchronization is complete, the backup data will be slightly behind the data in the master cache node. During instance backup, the standby cache node stops persisting the latest changes to disk files. If new data is written to the master cache node during backup, the backup file will not contain the new data. Backup time It is advisable to back up instance data during off-peak periods. Storage and pricing of backup files Backup files are stored to OBS. DCS provides the backup service free of charge, but OBS charges will be incurred for the amount and period that storage space is consumed. Handling exceptions in scheduled backup If a scheduled backup task is triggered while the DCS instance is restarting or being scaled up, the scheduled backup task will be run in the next cycle

40 3 Operation Guide Data Restoration If backing up a DCS instance fails or the backup is postponed because another task is in progress, DCS will try to back up the instance in the next cycle. A maximum of three retries are allowed within a single day. Retention period of backup data Scheduled backup files are retained for up to seven days. The retention period is user configurable. At the end of the retention period, most backup files of the DCS instance will be automatically deleted, but at least one backup file will be retained. Manual backup files are retained permanently and need to be manually deleted. Data restoration process a. A user initiates a data restoration request using the DCS console. b. DCS obtains the backup file from OBS. c. Read/write to the DCS instance is suspended. d. The original data persistence file of the master cache node is replaced by the backup file. e. The new data persistence file (that is, the backup file) is reloaded. f. Data is restored, and the DCS instance starts to provide read/write service again. Impact on service systems Data read/write is suspended during instance data restoration and resumed after data in the master cache node is restored. Handling data restoration exceptions If a backup file is corrupted, DCS will try to fix the backup file while restoring instance data. If the backup file is successfully fixed, DCS proceeds to restore instance data. If the backup file cannot be fixed, the DCS instance (if the instance type is master/standby) or the cache node with an unrepairable backup file (if the instance type is cluster) will be restored to the state in which it was before data restoration Configuring a Backup Policy Scenario Prerequisites Procedure On the DCS console, you can configure an automatic backup policy. The system then backs up data in your instances according to the backup policy. If automatic backup is not required, disable the automatic backup function in the backup policy. At least one master/standby DCS instance has been created. Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project

41 3 Operation Guide Step 3 Click Service List, and choose Database > to launch the DCS console. Step 4 Step 5 Step 6 Step 7 Step 8 Step 9 In the navigation pane, choose Cache Manager. On the Cache Manager page, filter DCS instances by instance status and/or name to find the DCS instance for which you want to configure a backup policy. Click the name of the chosen DCS instance to display more details about the DCS instance. On the instance details page, click Backup and Restore. On the Backup and Restore page, click Backup Policy. In the Backup Policy dialog box, set the Auto Backup parameter to On and specify other backup parameters. Click OK to complete setting the backup policy. Table 3-2 Parameters in a backup policy Parameter Auto Backup Backup Schedule Retention Period (days) Start Time Description An indicator of whether automatic backup is enabled. Day of a week on which data in the chosen DCS instance is automatically backed up. You can select one or multiple days of a week. The number of days that automatically backed up data is retained. Backup data will be permanently deleted at the end of retention period and cannot be restored. Data can be retained for up to 7 days. Time at which automatic backup starts. Value: the full hour between 00:00 to 23:00 The DCS checks backup policies once every hour. If the backup start time in a backup policy has arrived, data in the corresponding instance is backed up. NOTE Instance backup takes 5 to 30 minutes. The data added or modified during the backup process will not be backed up. To reduce the impact of backup on services, it is recommended that data should be backed up during off-peak periods. Only instances in the Running state can be backed up. ----End Manually Backing Up a DCS Instance Scenario On the DCS console, you can manually back up data in instances

42 3 Operation Guide By default, manually backed up data is permanently retained. If backup data is no longer in use, you can delete it manually. Prerequisites At least one master/standby DCS instance is in the Running state. Procedure Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Step 6 Step 7 Step 8 Step 9 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. On the Cache Manager page, filter DCS instances by instance status and/or name to find the DCS instance you want to manually back up. Click the name of the chosen DCS instance to display more details about the DCS instance. On the instance details page, click Backup and Restore. On the Backup and Restore page, click Manual Backup. In the Manual Backup dialog box, click OK to start manual backup. NOTE ----End Information in the Remarks text box cannot exceed 128 bytes. Instance backup takes 10 to 15 minutes. The data added or modified during the backup process will not be backed up Restoring a DCS Instance Scenario On the DCS console, you can restore backup data to a chosen DCS instance. Prerequisites At least one master/standby DCS instance is in the Running state. A backup task has been run to back up data in the instance to be restored and the status of the backup task is Successful. Procedure Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project

43 3 Operation Guide Step 3 Click Service List, and choose Database > to launch the DCS console. Step 4 Step 5 Step 6 Step 7 Step 8 Step 9 In the navigation pane, choose Cache Manager. On the Cache Manager page, filter DCS instances by instance status and/or name to find the DCS instance you want to restore. Click the name of the chosen DCS instance to display more details about the DCS instance. On the instance details page, click Backup and Restore. Click Restore in the same row as the chosen backup task. In the Restore DCS Instance dialog box, click OK to start instance restoration. NOTE ----End Information in the Description text box cannot exceed 128 bytes. The Restoration History tab page displays the result of the instance restoration task. Instance restoration takes 5 to 30 minutes. While being restored, DCS instances do not accept data operation requests from clients because existing data is being overwritten by the backup data

44 4 FAQs 4 FAQs 4.1 Does DCS Support Common Redis Clients Such as Jedis? DCS supports all clients compatible with Redis, including Jedis. You can download Redis clients from Should I Install Redis If I Already Have a DCS Instance? It is not necessary to install Redis in this case. You can access your DCS instance through a Redis client. 4.3 Do I Need to Enter a Password When Accessing a DCS Instance? Where Can I Obtain the Password? Redis allows clients to access cache service directly without any password. In DCS, additional efforts have been made to harden Redis. A client must enter the correct password before it is allowed to access DCS instances. The instance password is manually assigned when you are creating a DCS instance. 4.4 Does Standby Cache Node Remain Synchronized With Master Cache Node? Generally, updates to the master cache node are automatically and asynchronously replicated to the standby cache node. This means that data in the standby cache node may not always be consistent with data in the master cache node. This inconsistency is typically seen in either of the following situations: The input/output (I/O) write speed of the master cache node is faster than the synchronization speed of the standby cache node

45 4 FAQs There is network latency between the master and standby cache nodes. If a failover occurs when some data is not yet replicated to the standby cache node, the unreplicated data may be lost after the failover. 4.5 Does DCS Support Data Persistence? Only master/standby and cluster DCS instances support data persistence. 4.6 Can DCS Be Accessed over Public Networks? DCS cannot be accessed over public networks. The ECS serving as your Redis client and the DCS instances that the Redis client want to access must belong to the same VPC. 4.7 Can I Modify Configuration Parameters of DCS Instances? You can modify configuration parameters for DCS instances in the Running state. Note that configuration parameters of DCS instances in cluster mode are not modifiable. 4.8 Can I Recover Data from Deleted DCS Instances? Data that is automatically deleted by DCS instances or manually deleted by users using Redis clients cannot be recovered. By default, data is not evicted from DCS instances. However, you can modify the value of the maxmemory-policy parameter to adjust the eviction policy, and then DCS instances evict keys according to the eviction policy. If a DCS instance is deleted, instance data will also be removed and cannot be recovered. Therefore, exercise caution when deleting DCS instances. Master/Standby DCS instances support data backup. If you have backed up the instance data before deleting them, the deleted instance data can be recovered. 4.9 What Is the Default Data Eviction Policy? Data is evicted from the cache based on the user-defined space limit in order to make space for new data. By default, data is not evicted from DCS instances. In the current version of DCS, you can select an eviction policy. When maxmemory is reached, you can select one of the following six eviction policies: noeviction When the memory limit is reached, DCS instances return errors to clients and no longer process write requests or other requests that could result in more memory being used. However, DEL and a few more exception requests can continue to be processed. allkeys-lru DCS instances try to evict the least recently used keys first, in order to make space for new data

46 4 FAQs volatile-lru DCS instances try to evict the least recently used keys with an expire set first, in order to make space for new data. allkeys-random DCS instances evict random keys in order to make space for new data. volatile-random DCS instances evict random keys with an expire set, in order to make space for new data. volatile-ttl DCS instances evict keys with an expire set, and try to evict keys with a shorter time to live (TTL) first, in order to make space for new data. NOTE If the configured policy is volatile-lru, volatile-random, or volatile-ttl, and no keys match the configured policy, the eviction behavior will match noeviction What Should I Do If the Connection Between My Redis Client and the DCS Instance Times Out? Reconnect your client to the DCS instance. The timeout is probably due to network connectivity problems or DCS instance errors. For example, if your DCS instance is overloaded or experiences a failover, your Redis client returns the error message "Connection timed out" or "Connection reset by peer". Some Redis clients do not support automatic reconnection. If you use these Redis clients, the applications that use DCS must be able to proactively initiate reconnection Why Is Available Memory of Unused DCS Instances Less Than Total Memory and Why Is Memory Usage of Unused DCS Instances Greater Than Zero? Before a newly created DCS instance is put into use, the available memory is less than the total memory because some memory is reserved for system overhead and data persistence (supported by master/standby instances). For more information about the available memory for each type of DCS instance, see DCS Instance Specifications. DCS instances use a certain amount of memory for Redis-server buffers and internal data structures. This is why memory usage of unused DCS instances is greater than zero Does DCS Support Data Read/Write in Multiple Databases? Both single-node and master/standby DCS instances support data read/write in multiple databases. By default, single-node and master/standby DCS instances can read and write data in 16 databases (databases 0 15). DCS instances in cluster mode do not support data read/write in multiple databases

47 4 FAQs 4.13 Why Can't I Back Up or Scale Up My DCS Instances? Only single-node and master/standby DCS instances support specification scale-up, while cluster DCS instances do not. Only master/standby DCS instances support backup and restoration, while single-node and cluster DCS instances do not. DCS instance backup and scale-up are new functions provided by DCS. If your DCS instances were created long time ago, the version of these DCS instances is not upgraded and therefore cannot support some new functions (such as DCS instance backup and scale-up). You can contact technical support personnel to upgrade your DCS instance version. After the upgrade, you can back up and scale up your DCS instances Do DCS Instances in Cluster Mode Support Native Redis Clusters? DCS instances in cluster mode do not support native Redis clusters. DCS clusters use the proxy-based cluster mode, so when you query the parameters of the native Redis cluster, DCS retains the default values for some parameters, for example, cluster_enabled: Why Do My DCS Instances Sometimes Become Unavailable While They Are in Running State? The most probable cause is a network fault. To locate the fault, perform the following steps: Step 1 Log in to the management console. Step 2 Click in the upper left corner of the management console and select a region and a project. Step 3 Step 4 Step 5 Step 6 Click Service List, and choose Database > to launch the DCS console. In the navigation pane, choose Cache Manager. Click the name of the unavailable DCS instance. On the Summary tab page, click View next to Monitoring Metrics. View monitoring metrics of the DCS instance. If monitoring metrics change dramatically over a certain period of time, a network fault has occurred. After the network administrator rectifies the problem, reconnect to the DCS instance or perform management operations on the DCS instance again. If all monitoring metrics fall within the acceptable range, contact technical support. ----End

48 4 FAQs 4.16 Why Is the Used Instance Memory Displayed on Cloud Eye Console Slightly Higher than the Available Instance Memory? For DCS instances in single-node and master/standby modes, the used instance memory is measured by the redis-server process. For DCS instances in cluster mode, the used cluster memory is the sum of used memory of all shards in the cluster. The redis-server of each shard measures the used memory of the shard. Due to internal implementation of the open-source redis-server, the used instance memory is normally slightly higher than the available instance memory How to Select and Configure a Security Group? An ECS can communicate with a DCS instance if they belong to the same VPC and security group rules are configured correctly. It is recommended that the ECS and DCS instance belong to the same security group. After a security group is created, the security group rules, by default, include a rule allowing members in the security group to access each other without any restrictions. If the ECS and DCS instance belong to different security groups, add security group rules to ensure that the ECS and DCS instance can access each other. NOTE The example rules provided here are based on the assumption that: The ECS on which the client runs belongs to security group sg-ecs, and the DCS instance that the client will access belongs to security group sg-dcs. The port number of the DCS instance is The remote end is a security group or an IP address

49 4 FAQs Figure 4-1 Security group rules allowing the ECS to access the DCS instance Figure 4-2 Security group rules making the DCS instance accessible to the ECS 4.18 Do DCS Instances Support Cross-VPC Access? Generally, VPCs are isolated from each other and ECSs cannot access DCS instances that belong to a different VPC from these ECSs