---
title: Simulate Redis Faults
---

This document introduces how to use Chaosd to simulate Redis faults. This feature uses Golang interfaces in `go-redis` package and the `redis-server` command-line tool. You can create experiments either in command-line mode or service mode.

## Create experiments using command-line mode

Before creating an experiment, you can run the following command to see the Redis fault types that are supported by Chaosd:

```bash
chaosd attack redis -h
```

The result is as follows:

```bash
Redis attack related commands

Usage:
  chaosd attack redis [command]

Available Commands:
  cache-expiration  expire keys in Redis
  cache-limit       set maxmemory of Redis
  cache-penetration penetrate cache
  sentinel-restart  restart sentinel
  sentinel-stop     stop sentinel

Flags:
  -h, --help   help for redis

Global Flags:
      --log-level string   the log level of chaosd. The value can be 'debug', 'info', 'warn' and 'error'
      --uid string         the experiment ID

Use "chaosd attack redis [command] --help" for more information about a command.
```

Currently, Chaosd supports simulating cache expiration, cache penetration, cache limit, sentinel restart, and sentinel stop.

### Simulate cache expiration using command-line mode

The meaning of this command is the same as EXPIRE in Redis. For more details, refer to the [Redis official documentation](https://redis.io/commands/expire/).

:::note

Currently, Chaosd does not support recovering keys that have executed `cache-expiration`, so please backup them in advance if you want to recover them.

:::

#### Commands for cache expiration

```bash
chaosd attack redis cache-expiration -h
```

The result is as follows:

```bash
expire keys in Redis

Usage:
  chaosd attack redis cache-expiration [flags]

Flags:
  -a, --addr string         The address of redis server
      --expiration string   The expiration of the key. A expiration string should be able to be converted to a time duration, such as "5s" or "30m" (default "0")
  -h, --help                help for cache-expiration
  -k, --key string          The key to be set a expiration, default expire all keys
      --option string       The additional options of expiration, only NX, XX, GT, LT supported
  -p, --password string     The password of server

Global Flags:
      --log-level string   the log level of chaosd. The value can be 'debug', 'info', 'warn' and 'error'
      --uid string         the experiment ID
```

#### Configuration description for cache expiration

| Configuration item | Abbreviation | Type | Description | Value |
| :-- | :-- | :-- | :-- | :-- |
| `addr` | a | string | The address and port of Redis server to be injected into the fault, for example `127.0.0.1:6379` | Default value: `""` |
| `expiration` | None | string | The specified key will be expired after `expiration` arrives | Default value: `"0"`. Make sure that the string is in the format supported by `time.Duration` |
| `key` | k | string | The key to be expired | Default value: `""`, which means the expiration is set for all keys |
| `option` | None | string | Additional options for `expiration`. **Only versions of Redis after 7.0.0 support this flag** | Default value: `""`. Only NX, XX, GT, and LT are supported |
| `password` | p | string | The password to log in to the server | Default value: `""` |

#### Example for simulating cache expiration

```bash
chaosd attack redis cache-expiration -a 127.0.0.1:6379 --option GT --expiration 1m
```

### Simulate cache limit using command-line mode

#### Commands for cache limit

```bash
chaosd attack redis cache-limit -h
```

The result is as follows:

```bash
set maxmemory of Redis

Usage:
  chaosd attack redis cache-limit [flags]

Flags:
  -a, --addr string       The address of redis server
  -h, --help              help for cache-limit
  -p, --password string   The password of server
      --percent string    The percentage of maxmemory
  -s, --size string       The size of cache (default "0")

Global Flags:
      --log-level string   the log level of chaosd. The value can be 'debug', 'info', 'warn' and 'error'
      --uid string         the experiment ID
```

#### Configuration description for cache limit

| Configuration item | Abbreviation | Type | Description | Value |
| :-- | :-- | :-- | :-- | :-- |
| `addr` | a | string | The address and port of Redis server to be injected into the fault, such as `127.0.0.1:6379` | Default value: `""` |
| `password` | p | string | The password to log in to the server | Default value: `""` |
| `percent` | None | string | Specifies `maxmemory` as a percentage of the original value | Default value: `""` |
| `size` | s | string | Specifies the size of `maxmemory` | Default `0`, which means no limitation of memory |

#### Example for simulating cache limit

```bash
chaosd attack redis cache-limit -a 127.0.0.1:6379 -s 256M
```

### Simulate cache penetration using command-line mode

This command will send the specified number of `GET` requests to Redis server as quickly as possible using Redis Pipeline. Since the requested keys do not exist on the Redis server, these requests will cause a cache penetration phenomenon.

#### Commands for cache penetration

```bash
chaosd attack redis cache-penetration -h
```

The result is as follows:

```bash
penetrate cache

Usage:
  chaosd attack redis cache-penetration [flags]

Flags:
  -a, --addr string       The address of redis server
  -h, --help              help for cache-penetration
  -p, --password string   The password of server
      --request-num int   The number of requests

Global Flags:
      --log-level string   the log level of chaosd. The value can be 'debug', 'info', 'warn' and 'error'
      --uid string         the experiment ID
```

#### Configuration description for cache penetration

| Configuration item | Abbreviation | Type | Description | Value |
| :-- | :-- | :-- | :-- | :-- |
| `addr` | a | string | The address and port of Redis server to be injected into the fault, such as `127.0.0.1:6379` | Default value: `""` |
| `password` | p | string | The password to log in to the server | Default value: `""` |
| `request-num` | None | int | Specifies the number of requests to be sent to the Redis server | Default value: `0` |

#### Example for simulating cache penetration

```bash
chaosd attack redis cache-penetration -a 127.0.0.1:6379 --request-num 100000
```

### Simulate Sentinel restart using command-line mode

#### Commands for Sentinel restart

```bash
chaosd attack redis sentinel-restart -h
```

The result is as follows:

```bash
restart sentinel

Usage:
  chaosd attack redis sentinel-restart [flags]

Flags:
  -a, --addr string         The address of redis server
  -c, --conf string         The config of Redis server
      --flush-config         Force Sentinel to rewrite its configuration on disk (default true)
  -h, --help                help for sentinel-restart
  -p, --password string     The password of server
      --redis-path string   The path of the redis-server command

Global Flags:
      --log-level string   the log level of chaosd. The value can be 'debug', 'info', 'warn' and 'error'
      --uid string         the experiment ID
```

#### Configuration description for Sentinel restart

| Configuration item | Abbreviation | Type | Description | Value |
| :-- | :-- | :-- | :-- | :-- |
| `addr` | a | string | The address and port of Sentinel to be injected into the fault, such as `127.0.0.1:26379` | Default value: `""` |
| `conf` | c | string | Specifies the path of Sentinel config file, this file will be used to revover the Sentinel | Default value: `""` |
| `flush-config` | None | bool | Forces Sentinel to rewrite its configuration on disk, including the current Sentinel state | Default value: `true` |
| `password` | p | string | The password to log in to the server | Default value: `""` |
| `redis-path` | None | string | Specifies the path of `redis-server` command-line tool | Default value: `""` |

#### Example for simulating Sentinel restart

```bash
chaosd attack redis sentinel-restart -a 127.0.0.1:26379 --conf /home/redis-test/sentinel-26379.conf
```

### Simulate Sentinel stop using command-line mode

#### Commands for Sentinel stop

```bash
chaosd attack redis sentinel-stop -h
```

The result is as follows:

```bash
stop sentinel

Usage:
  chaosd attack redis sentinel-stop [flags]

Flags:
  -a, --addr string         The address of redis server
  -c, --conf string         The config path of Redis server
      --flush-config        Force Sentinel to rewrite its configuration on disk (default true)
  -h, --help                help for sentinel-stop
  -p, --password string     The password of server
      --redis-path string   The path of the redis-server command

Global Flags:
      --log-level string   the log level of chaosd. The value can be 'debug', 'info', 'warn' and 'error'
      --uid string         the experiment ID
```

#### Configuration description for Sentinel stop

| Configuration item | Abbreviation | Type | Description | Value |
| :-- | :-- | :-- | :-- | :-- |
| `addr` | a | string | The address and port of Sentinel to be injected into the fault, such as `127.0.0.1:26379` | Default value: `""` |
| `conf` | c | string | Specifies the path of Sentinel configuration file, which is used to recover the Sentinel | Default value: `""` |
| `flush-config` | None | bool | Forces Sentinel to rewrite its configuration on disk, including the current Sentinel state | Default value: `true` |
| `password` | p | string | The password to log in to the server | Default value: `""` |
| `redis-path` | None | string | Specifies the path of `redis-server` command-line tool | Default value: `""` |

#### Example for simulating Sentinel restart

```bash
chaosd attack redis sentinel-stop -a 127.0.0.1:26379 --conf /home/redis-test/sentinel-26379.conf
```

## Create Redis fault experiments using service mode

To create experiments using the service mode, follow the instructions below:

1. Run Chaosd in the service mode:

   ```bash
   chaosd server --port 31767
   ```

2. Send a `POST` HTTP request to the `/api/attack/redis` path of the Chaosd service.

   ```bash
   curl -X POST 127.0.0.1:31767/api/attack/redis -H "Content-Type:application/json" -d '{fault-configuration}'
   ```

   In the above command, you need to configure `fault-configuration` according to the fault types. For the corresponding parameters, refer to the parameters and examples of each fault type in the following sections.

:::note

When running an experiment, remember to record the UID of the experiment. When you want to end the experiment corresponding to the UID, you need to send a `DELETE` HTTP request to the `/api/attack/{uid}` path of the Chaosd service.

:::

### Simulate cache expiration using service mode

#### Parameters for simulating cache expiration

| Parameter | Description | Type | Value |
| :-- | :-- | :-- | :-- |
| `action` | Action of the experiment | string | set to "corrupt" |
| `addr` | The address and port of Redis server to be injected into the fault, such as `127.0.0.1:6379` | string | Default value: `""` |
| `expiration` | The specified key will be expired after `expiration` arrives | string | Default value: `"0"`. Make sure that the string is in the format supported by `time.Duration` |
| `key` | The key to be expired | string | Default value: `""`, which means the expiration is set for all keys |
| `option` | Additional options for `expiration`. **Only versions of Redis after 7.0.0 support this flag** | string | Default value: `""`. Only NX, XX, GT, and LT are supported |
| `password` | The password to log in to the server | Default value: `""` |

#### Example for simulating cache expiration using service mode

```bash
curl -X POST 127.0.0.1:31767/api/attack/redis -H "Content-Type:application/json" -d '{"action":"expiration", "expiration":"1m","addr":"127.0.0.1:6379"}'
```

### Simulate cache limit using service mode

#### Parameters for simulating cache limit

| Parameter | Description | Type | Value |
| :-- | :-- | :-- | :-- |
| `action` | Action of the experiment | string | set to "cacheLimit" |
| `addr` | The address and port of Redis server to be injected into the fault, such as `127.0.0.1:6379` | string | Default value: `""` |
| `password` | The password to log in to the server | string | Default value: `""` |
| `percent` | Specifies `maxmemory` as a percentage of the original value | string | Default value: `""` |
| `size` | Specifies the size of `maxmemory` | string | Default `0`, which means no limitation of memory |

#### Example for simulating cache limit using service mode

```bash
curl -X POST 127.0.0.1:31767/api/attack/redis -H "Content-Type:application/json" -d '{"action":"cacheLimit", ""addr":"127.0.0.1:6379", "percent":"50%"}'
```

### Simulate cache penetration using service mode

#### Parameters for simulating cache penetration

| Parameter | Description | Type | Value |
| :-- | :-- | :-- | :-- |
| `action` | Action of the experiment | string | set to "penetration" |
| `addr` | The address and port of Redis server to be injected into the fault, such as `127.0.0.1:6379` | string | Default value: `""` |
| `password` | The password to log in to the server | string | Default value: `""` |
| `request-num` | Specifies the number of requests to be sent to the Redis server | int | Default value: `0` |

#### Example for simulating cache penetration using service mode

```bash
curl -X POST 127.0.0.1:31767/api/attack/redis -H "Content-Type:application/json" -d '{"action":"penetration", ""addr":"127.0.0.1:6379", "request-num":"10000"}'
```

### Simulate Sentinel restart using service mode

#### Parameters for simulating Sentinel restart

| Parameter | Description | Type | Value |
| :-- | :-- | :-- | :-- |
| `action` | Action of the experiment | string | set to "restart" |
| `addr` | The address and port of Sentinel to be injected into the fault, such as `127.0.0.1:26379` | string | Default value: `""` |
| `conf` | Specifies the path of Sentinel configuration file, which is used to recover the Sentinel | string | Default value: `""` |
| `flush-config` | Forces Sentinel to rewrite its configuration on disk, including the current Sentinel state | bool | Default value: `true` |
| `password` | The password to log in to the server | string | Default value: `""` |
| `redis-path` | Specifies the path of `redis-server` command-line tool | string | Default value: `""` |

#### Example for simulating Sentinel restart using service mode

```bash
curl -X POST 127.0.0.1:31767/api/attack/redis -H "Content-Type:application/json" -d '{"action":"restart", ""addr":"127.0.0.1:26379", "conf":"/home/redis-test/sentinel-26379.conf"}'
```

### Simulate Sentinel stop using service mode

#### Parameters for simulating Sentinel stop

| Parameter | Description | Type | Value |
| :-- | :-- | :-- | :-- |
| `action` | Action of the experiment | string | set to "stop" |
| `addr` | The address and port of Sentinel to be injected into the fault, such as `127.0.0.1:26379` | string | Default value: `""` |
| `conf` | Specifies the path of Sentinel configuration file, which is used to recover the Sentinel | string | Default value: `""` |
| `flush-config` | Forces Sentinel to rewrite its configuration on disk, including the current Sentinel state | bool | Default value: `true` |
| `password` | The password to log in to the server | string | Default value: `""` |
| `redis-path` | Specifies the path of `redis-server` command-line tool | string | Default value: `""` |

#### Example for simulating Sentinel stop using service mode

```bash
curl -X POST 127.0.0.1:31767/api/attack/redis -H "Content-Type:application/json" -d '{"action":"stop", ""addr":"127.0.0.1:26379", "conf":"/home/redis-test/sentinel-26379.conf"}'
```