Add the cdc readme and usage doc

Signed-off-by: SimFG <bang.fu@zilliz.com>

Makefile

@@ -1,5 +1,8 @@
 PWD 	  := $(shell pwd)
 
+build:
+	$(MAKE) -C server build
+
 test-go:
 	@echo "Running go unittests..."
 	@(env bash $(PWD)/scripts/run_go_unittest.sh)

README.md

@@ -1,3 +1,41 @@
 # Milvus-CDC
 
-Milvus-CDC is a change data capture tool for Milvus. It can capture the changes of upstream Milvus collections and sink them to downstream Milvus.
+CDC is "Change Data Capture", and Milvus-CDC is a change data capture tool for Milvus. It can capture the changes of upstream Milvus collections and sink them to downstream Milvus. This will bring the following benefits:
+
+1. Data reliability is improved and the probability of data loss is reduced;
+2. Based on CDC, the active-standby disaster recovery feature of milvus can be implemented to ensure that even if the milvus-source cluster fails, it can be quickly switched to ensure the availability of upper-layer services;
+
+## Quick Start
+
+**Please use the source code to compile and use it now**, because milvus-cdc is undergoing rapid iteration, the image currently provided is not stable, and the implementation method has also changed a lot.
+
+```bash
+git clone https://github.com/zilliztech/milvus-cdc.git
+
+make build
+```
+
+After successfully building, the `cdc` bin file will be generated in the `server` directory.
+
+**DON'T execute it directly.** If you do it, I think you must get a error because you need to configure it before using it. How to configure and use cdc, refer to: [Milvus-CDC Usage](doc/cdc-usage.md)
+
+## Basic Components
+
+At present, cdc mainly consists of two parts: http server and corelib.
+
+- The http server, is responsible for accepting user-side requests, controlling task execution, and maintaining meta-information;
+- corelib,  is responsible for synchronizing the execution of tasks, including reader and writer:
+   - reader reads relevant information from etcd and mq of source Milvus;
+   - The writer converts the msg in mq into Milvus api parameters and sends the request to the target Milvus;
+
+![components](doc/pic/milvus-cdc-components.png)
+
+## CDC Data Processing Flow
+
+1. User creates cdc task through http interface;
+2. Obtain collection-related meta-information through etcd in Milvus-source, such as the channel information and checkpoint information corresponding to the collection, etc;
+3. After obtaining the meta-information related to the collection, connect to mq(message queue) to subscribe to the data;
+4. Read the data in mq, parse the data and forward it through go-sdk or perform the same operation as milvus-source;
+
+![flow](doc/pic/milvus-cdc-data.png)
+

doc/cdc-usage.md

@@ -0,0 +1,291 @@
+# Milvus-CDC Usage
+
+## Limitation
+
+1. Only the cluster dimension can be synchronized, that is, the collection name can only be specified as `*` when creating a task. The collection dimension will also be supported in the future.
+2. Supported operations
+- Create/Drop Collection
+- Insert/Delete
+- Create/Drop Partition
+- Create/Drop Index
+- Load/Release/Flush
+- Create/Drop Database
+
+    **Anything not mentioned is not supported;**
+
+3. Milvus cdc only supports synchronized data. If you need active and backup disaster recovery functions, please contact us;
+
+## Configuration
+
+The configuration consists of two main parts: the Milvus configuration for Milvus-Target and the CDC startup configuration.
+
+Furthermore, please ensure that both the source and target Milvus versions are **2.3.2 or higher**. Any other versions are not supported.
+
+1. Milvus-Target Config
+
+    Set the value of `common.ttMsgEnabled` in the milvus.yaml file in the milvus-target cluster to `false`.
+
+2. CDC Startup Config
+
+    The following is the cdc startup configuration file template. You can see this file in the `server/configs` directory.
+
+```yaml
+# cdc server address
+address: 0.0.0.0:8444
+# max task num
+maxTaskNum: 100
+# max task name length
+maxNameLength: 256
+
+# cdc meta data config
+metaStoreConfig:
+  # the metastore type, available value: etcd, mysql
+  storeType: etcd
+  # etcd address
+  etcdEndpoints:
+    - localhost:2379
+  # mysql connection address
+  mysqlSourceUrl: root:root@tcp(127.0.0.1:3306)/milvus-cdc?charset=utf8
+  # meta data prefix, if multiple cdc services use the same store service, you can set different rootPaths to achieve multi-tenancy
+  rootPath: cdc
+
+# milvus-source config, these settings are basically the same as the corresponding configuration of milvus.yaml in milvus source.
+sourceConfig:
+  # etcd config
+  etcdAddress:
+    - localhost:2379
+  etcdRootPath: by-dev
+  etcdMetaSubPath: meta
+  # default partition name
+  defaultPartitionName: _default
+  # read buffer length, mainly used for buffering if writing data to milvus-target is slow.
+  readChanLen: 10
+  # milvus-target mq config, which is pulsar or kafka
+  pulsar:
+    address: pulsar://localhost:6650
+    webAddress: localhost:80
+    maxMessageSize: 5242880
+    tenant: public
+    namespace: default
+#    authPlugin: org.apache.pulsar.client.impl.auth.AuthenticationToken
+#    authParams: token:xxx
+#  kafka:
+#    address: 127.0.0.1:9092
+```
+
+After completing these two steps of configuration, do not rush to start CDC. Need to ensure the following points:
+
+1. Whether the `common.ttMsgEnabled` configuration value of milvus-target is `false``;
+2. Confirm that the `mq` type configured by cdc is the same as the type configured by milvus-source;
+3. Ensure that the network environment where the cdc service is located can correctly connect to the mq and etcd addresses in the milvus-source in the configuration;
+
+## Usage
+
+All http APIs need to comply with the following rules:
+
+- Request method: POST
+- Request path:/cdc
+- Request body:
+
+```json
+{
+     "request_type": "",
+     "request_data": {}
+}
+```
+
+Different requests are distinguished by `request_type`, which currently includes: create, delete, pause, resume, get and list. If the request fails, a non-200 http status code will be returned.
+
+### create request
+
+- milvus_connect_param, the connection params of the milvus-target server;
+- collection_infos, the collection information that needs to be synchronized, which currently only supports `*`;
+- rpc_channel_info, the corresponding name value is composed of the two values ​​of `common.chanNamePrefix.cluster` and `common.chanNamePrefix.replicateMsg` in **milvus-source**, connected by the symbol `-`
+
+```http
+POST localhost:8444/cdc
+Content-Type: application/json
+
+body:
+{
+  "request_type":"create",
+  "request_data":{
+    "milvus_connect_param":{
+      "host":"localhost",
+      "port":19530,
+      "username":"root",
+      "password":"Milvus",
+      "enable_tls":true,
+      "connect_timeout":10
+    },
+    "collection_infos":[
+      {
+        "name":"*"
+      }
+    ],
+    "rpc_channel_info": {
+      "name": "by-dev-replicate-msg"
+    }
+  }
+}
+```
+
+After success, the task_id will be returned, such as:
+
+```json
+{"task_id":"6623ae52d35842a5a2c9d89b16ed7aa1"}
+```
+
+If there is an exception, an http error will appear.
+
+### delete request
+
+delete a cdc task.
+
+**request**
+
+```http
+POST localhost:8444/cdc
+Content-Type: application/json
+
+body:
+{
+  "request_type":"delete",
+  "request_data": {
+    "task_id": "f84605ae48fb4170990ab80addcbd71e"
+  }
+}
+```
+
+**response**
+
+```json
+{}
+```
+
+### pause request
+
+pause a cdc task.
+
+**request**
+
+```http
+POST localhost:8444/cdc
+Content-Type: application/json
+
+body:
+{
+  "request_type":"pause",
+  "request_data": {
+    "task_id": "4d458a58b0f74e85b842b1244dc69546"
+  }
+}
+```
+
+**response**
+
+```json
+{}
+```
+
+### resume request
+
+resume a cdc task.
+
+**request**
+
+```http
+POST localhost:8444/cdc
+Content-Type: application/json
+
+body:
+{
+  "request_type":"resume",
+  "request_data": {
+    "task_id": "4d458a58b0f74e85b842b1244dc69546"
+  }
+}
+```
+
+**response**
+
+```json
+{}
+```
+
+### get request
+
+get a cdc task info
+
+**request**
+
+```http
+POST localhost:8444/cdc
+Content-Type: application/json
+
+body:
+{
+  "request_type":"get",
+  "request_data": {
+    "task_id": "4d458a58b0f74e85b842b1244dc69546"
+  }
+}
+```
+
+**response**
+
+```json
+{
+    "task_id":"4d458a58b0f74e85b842b1244dc69546",
+    "Milvus_connect_param":{
+        "host":"localhost",
+        "port":19530,
+        "connect_timeout":10
+    },
+    "collection_infos":[
+        {
+            "name":"*"
+        }
+    ],
+    "state":"Running"
+}
+```
+
+### list request
+
+list the info of all cdc tasks.
+
+**request**
+
+```http
+POST localhost:8444/cdc
+Content-Type: application/json
+
+body:
+{
+  "request_type":"list"
+}
+```
+
+**response**
+
+```json
+{
+    "tasks": [
+        {
+            "task_id": "4d458a58b0f74e85b842b1244dc69546",
+            "Milvus_connect_param": {
+                "host": "localhost",
+                "port": 19530,
+                "connect_timeout": 10
+            },
+            "collection_infos": [
+                {
+                    "name": "*"
+                }
+            ],
+            "state": "Running"
+        }
+    ]
+}
+```

server/configs/cdc.yaml

@@ -23,16 +23,4 @@ sourceConfig:
 #    authParams: token:xxx
 #  kafka:
 #    address: 127.0.0.1:9092
-enableReverse: false
-reverseMilvus:
-  host: localhost
-  port: 19530
-  username: root
-  password: 123456
-  enable_tls: false
-  ignore_partition: true
-  connect_timeout: 10
-currentMilvus:
-  host: 127.0.0.1
-  port: 19530
 maxNameLength: 256