天天看點

elasticsearch-jdbc實作MySQL同步到ElasticSearch深入詳解

elasticsearch-jdbc實作MySQL同步到ElasticSearch深入詳解
點此連結看原視訊

1.如何實作mysql與elasticsearch的資料同步?

逐條轉換為json顯然不合适,需要借助第三方工具或者自己實作。核心功能點:同步增、删、改、查同步。

2、mysql與elasticsearch同步的方法有哪些?優缺點對比?

目前該領域比較牛的插件有:

1)、elasticsearch-jdbc,嚴格意義上它已經不是第三方插件。已經成為獨立的第三方工具。

https://github.com/jprante/elasticsearch-jdbc

2)、elasticsearch-river-mysql插件

https://github.com/scharron/elasticsearch-river-mysql

3)、go-mysql-elasticsearch(國内作者siddontang)

https://github.com/siddontang/go-mysql-elasticsearch

1-3同步工具/插件對比:

go-mysql-elasticsearch仍處理開發不穩定階段。

為什麼選擇elasticsearch-jdbc而不是elasticsearch-river-mysql插件的原因?(參考:

http://stackoverflow.com/questions/23658534/using-elasticsearch-river-mysql-to-stream-data-from-mysql-database-to-elasticsea

1)通用性角度:elasticsearch-jdbc更通用,

2)版本更新角度:elasticsearch-jdbc GitHub活躍度很高,最新的版本2.3.3.02016年5月28日相容Elasticsearch2.3.3版本。

而elasticsearch-river-mysql 2012年12月13日後便不再更新。

綜上,選擇elasticsearch-jdbc作為mysql同步Elasticsearch的工具理所當然。

elasticsearch-jdbc的缺點與不足(他山之石):

1)、go-mysql-elasticsearch作者siddontang在部落格提到的:

elasticsearch-river-jdbc的功能是很強大,但并沒有很好的支援增量資料更新的問題,它需要對應的表隻增不減,而這個幾乎在項目中是不可能辦到的。

http://www.jianshu.com/p/05cff717563c

2)、

部落客leotse90在博文中提到elasticsearch-jdbc的缺點:那就是删除操作不能同步(實體删除)!

http://leotse90.com/2015/11/11/ElasticSearch

與MySQL資料同步以及修改表結構/

我截止2016年6月16日沒有測試到,不妄加評論。

elasticsearch-jdbc實作MySQL同步到ElasticSearch深入詳解

3、elasticsearch-jdbc如何使用?要不要安裝?

3.1 和早期版本不同點

elasticsearch-jdbcV2.3.2.0版本不需要安裝。以下筆者使用的elasticsearch也是2.3.2測試。

作業系統:CentOS release 6.6 (Final)

看到這裡,你可能會問早期的版本有什麼不同呢?很大不同。從我搜集資料來看,不同點如下:

1)早期1.x版本,作為插件,需要安裝。

2)配置也會有不同。

3.2 elasticsearch-jdbc使用(同步方法一)

前提:

1)elasticsearch 2.3.2 安裝成功,測試ok。

2)mysql安裝成功,能實作增、删、改、查。

可供測試的資料庫為test,表為cc,具體資訊如下:

mysql> select * from cc;

+----+------------+

| id | name |

| 1 | laoyang |

| 2 | dluzhang |

| 3 | dlulaoyang |

3 rows in set (0.00 sec)

1

2

3

4

5

6

7

8

9

第一步:下載下傳工具。

址:

http://xbib.org/repository/org/xbib/elasticsearch/importer/elasticsearch-jdbc/2.3.2.0/elasticsearch-jdbc-2.3.2.0-dist.zip

第二步:導入Centos。路徑自己定,筆者放到根目錄下,解壓。unzip elasticsearch-jdbc-2.3.2.0-dist.zip

第三步:設定環境變量。

[root@5b9dbaaa148a /]# vi /etc/profile

export JDBC_IMPORTER_HOME=/elasticsearch-jdbc-2.3.2.0

使環境變量生效:

[root@5b9dbaaa148a /]# source /etc/profile

第四步:配置使用。詳細參考:

1)、根目錄下建立檔案夾odbc_es 如下:

[root@5b9dbaaa148a /]# ll /odbc_es/

drwxr-xr-x 2 root root 4096 Jun 16 03:11 logs

-rwxrwxrwx 1 root root 542 Jun 16 04:03 mysql_import_es.sh

2)、建立腳本mysql_import_es.sh,内容如下;

[root@5b9dbaaa148a odbc_es]# cat mysql_import_es.sh

’#!/bin/sh

bin=$JDBC_IMPORTER_HOME/bin

lib=$JDBC_IMPORTER_HOME/lib

echo '{

"type" : "jdbc",

"jdbc": {

"elasticsearch.autodiscover":true,

"elasticsearch.cluster":"my-application", #簇名,詳見:/usr/local/elasticsearch/config/elasticsearch.yml

"url":"jdbc:mysql://10.8.5.101:3306/test", #mysql資料庫位址

"user":"root", #mysql使用者名

"password":"123456", #mysql密碼

"sql":"select * from cc",

"elasticsearch" : {

 "host" : "10.8.5.101",

 "port" : 9300

},

"index" : "myindex", #新的index

"type" : "mytype" #新的type

}

}'| java \

 -cp "${lib}/*" \

 -Dlog4j.configurationFile=${bin}/log4j2.xml \

 org.xbib.tools.Runner \

 org.xbib.tools.JDBCImporter

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

3)、為 mysql_import_es.sh 添加可執行權限。

[root@5b9dbaaa148a odbc_es]# chmod a+x mysql_import_es.sh

4)執行腳本mysql_import_es.sh

[root@5b9dbaaa148a odbc_es]# ./mysql_import_es.sh

第五步:測試資料同步是否成功。

使用elasticsearch檢索查詢:

[root@5b9dbaaa148a odbc_es]# curl -XGET 'http://10.8.5.101:9200/myindex/mytype/_search?pretty'

{

 "took" : 4,

 "timed_out" : false,

 "_shards" : {

 "total" : 8,

 "successful" : 8,

 "failed" : 0

 },

 "hits" : {

 "total" : 3,

 "max_score" : 1.0,

 "hits" : [ {

 "_index" : "myindex",

 "_type" : "mytype",

 "_id" : "AVVXKgeEun6ksbtikOWH",

 "_score" : 1.0,

 "_source" : {

 "id" : 1,

 "name" : "laoyang"

 }

 }, {

 "_id" : "AVVXKgeEun6ksbtikOWI",

 "id" : 2,

 "name" : "dluzhang"

 "_id" : "AVVXKgeEun6ksbtikOWJ",

 "id" : 3,

 "name" : "dlulaoyang"

 } ]

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

出現以上包含mysql資料字段的資訊則為同步成功。

4、 elasticsearch-jdbc 同步方法二

[root@5b9dbaaa148a odbc_es]# cat mysql_import_es_simple.sh

#!/bin/sh

 java \

 org.xbib.tools.JDBCImporter statefile.json

[root@5b9dbaaa148a odbc_es]# cat statefile.json

"elasticsearch.cluster":"my-application",

"url":"jdbc:mysql://10.8.5.101:3306/test",

"user":"root",

"password":"123456",

"index" : "myindex_2",

"type" : "mytype_2"

腳本和json檔案分開,腳本執行前先加載json檔案。

執行方式:直接運作腳本 ./mysql_import_es_simple.sh 即可。

5、Mysql與elasticsearch等價查詢

目标:實作從表cc中查詢id=3的name資訊。

1)MySQL中sql語句查詢:

mysql> select * from cc where id=3;

1 row in set (0.00 sec)

2)elasticsearch檢索:

[root@5b9dbaaa148a odbc_es]# curl

http://10.8.5.101:9200/myindex/mytype/_search?pretty

-d '

"filter" : { "term" : { "id" : "3" } }

}'

 "took" : 3,

 "total" : 1,

常見錯誤:

錯誤日志位置:/odbc_es/logs

日志内容:

[root@5b9dbaaa148a logs]# tail -f jdbc.log

[04:03:39,570][INFO ][org.xbib.elasticsearch.helper.client.BaseTransportClient][pool-3-thread-1] after auto-discovery connected to [{5b9dbaaa148a}{aksn2ErNRlWjUECnp_8JmA}{10.8.5.101}{10.8.5.101:9300}{master=true}]

Bug1、[02:46:23,894][ERROR][importer.jdbc ][pool-3-thread-1] error while processing request: cluster state is RED and not YELLOW, from here on, everything will fail!

原因:

you created an index with replicas but you had only one node in the cluster. One way to solve this problem is by allocating them on a second node. Another way is by turning replicas off.

你建立了帶副本 replicas 的索引,但是在你的簇中隻有一個節點。

解決方案:

方案一:允許配置設定‘它們’到第二個節點。

方案二:關閉副本replicas(非常可行)。如下:

curl -XPUT 'localhost:9200/_settings' -d '

 "index" : {

 "number_of_replicas" : 0

Bug2、[13:00:37,137][ERROR][importer.jdbc ][pool-3-thread-1] error while processing request: no cluster nodes available, check settings {autodiscover=false, client.transport.ignore_cluster_name=false, client.transport.nodes_sampler_interval=5s, client.transport.ping_timeout=5s, cluster.name=elasticsearch,

org.elasticsearch.client.transport.NoNodeAvailableException: no cluster nodes available, check

見上腳本中新增:

“elasticsearch.cluster”:“my-application”, #簇名,和/usr/local/elasticsearch/config/elasticsearch.yml 簇名保持一緻。

參考:

http://stackoverflow.com/questions/11944915/getting-an-elasticsearch-cluster-to-green-cluster-setup-on-os-x