前言

Elasticsearch 是一个开源的分布式 RESTful 搜索和分析引擎。它可以在近实时条件下，存储，查询和分析海量的数据。它还支持将快照备份至HDFS/S3上面，而阿里云OSS兼容S3的API，本文将介绍如何使用ES的Repository-S3插件将快照备份至OSS。

部署与配置

首先，我们需要安装repository-s3，可以参考官方文档：

https://www.elastic.co/guide/en/elasticsearch/plugins/7.2/repository-s3.html

启动ES，我们可以从log中看到，ES已经load了这个plugin：

[2019-07-15T14:12:09,225][INFO ][o.e.p.PluginsService ] [master] loaded module [aggs-matrix-stats]
[2019-07-15T14:12:09,225][INFO ][o.e.p.PluginsService ] [master] loaded module [analysis-common]
[2019-07-15T14:12:09,225][INFO ][o.e.p.PluginsService ] [master] loaded module [ingest-common]
[2019-07-15T14:12:09,226][INFO ][o.e.p.PluginsService ] [master] loaded module [ingest-geoip]
[2019-07-15T14:12:09,226][INFO ][o.e.p.PluginsService ] [master] loaded module [ingest-user-agent]
[2019-07-15T14:12:09,226][INFO ][o.e.p.PluginsService ] [master] loaded module [lang-expression]
[2019-07-15T14:12:09,226][INFO ][o.e.p.PluginsService ] [master] loaded module [lang-mustache]
[2019-07-15T14:12:09,227][INFO ][o.e.p.PluginsService ] [master] loaded module [lang-painless]
[2019-07-15T14:12:09,227][INFO ][o.e.p.PluginsService ] [master] loaded module [mapper-extras]
[2019-07-15T14:12:09,227][INFO ][o.e.p.PluginsService ] [master] loaded module [parent-join]
[2019-07-15T14:12:09,227][INFO ][o.e.p.PluginsService ] [master] loaded module [percolator]
[2019-07-15T14:12:09,227][INFO ][o.e.p.PluginsService ] [master] loaded module [rank-eval]
[2019-07-15T14:12:09,228][INFO ][o.e.p.PluginsService ] [master] loaded module [reindex]
[2019-07-15T14:12:09,228][INFO ][o.e.p.PluginsService ] [master] loaded module [repository-url]
[2019-07-15T14:12:09,228][INFO ][o.e.p.PluginsService ] [master] loaded module [transport-netty4]
[2019-07-15T14:12:09,228][INFO ][o.e.p.PluginsService ] [master] loaded plugin [repository-s3]
[2019-07-15T14:12:12,375][INFO ][o.e.d.DiscoveryModule ] [master] using discovery type [zen] and seed hosts providers [settings]
[2019-07-15T14:12:12,801][INFO ][o.e.n.Node ] [master] initialized
[2019-07-15T14:12:12,802][INFO ][o.e.n.Node ] [master] starting ...

然后，我们需要将OSS使用的Access Key和Secret Key配置到ES去，分别执行下面的命令：

bin/elasticsearch-keystore add s3.client.default.access_key
bin/elasticsearch-keystore add s3.client.default.secret_key

运行

首先，我们创建一个备份：

[root@master ~]# curl -XPUT 'http://localhost:9200/_snapshot/test' -H 'Content-Type: application/json' -d '{ "type": "s3", "settings": { "bucket": "hadoop-oss-test", "endpoint": "oss-cn-zhangjiakou-internal.aliyuncs.com"} }'
{"acknowledged":true}

NOTE: 上面的命令默认使用https协议来传输数据，如果想使用http协议，需要将"protocol": "http", "disable_chunked_encoding": true加到settings里面(这个特性将会在新版本发布后可用)。

可以使用下面的命令来确实创建是否成功：

[root@master ~]# curl -XGET localhost:9200/_snapshot/test?pretty
{
 "test" : {
 "type" : "s3",
 "settings" : {
 "bucket" : "hadoop-oss-test",
 "endpoint" : "oss-cn-zhangjiakou-internal.aliyuncs.com"
 }
 }
}

我们可以写入一些测试数据到ES，然后看下目前集群的索引信息:

[root@master ~]# curl -X GET "localhost:9200/_cat/indices?v"
health status index uuid pri rep docs.count docs.deleted store.size pri.store.size
green open sales 89ouBy6RQsuT34QRbn_jeQ 10 0 271786 0 15mb 15mb
green open customer fQCMEvXsQOu0UgMm1SAJlA 5 0 10000 0 717kb 717kb

假设我们只备份sales索引：

[root@master ~]# curl -XPUT 'http://localhost:9200/_snapshot/test/sales' -H 'Content-Type: application/json' -d '{ "indices": "sales" }'
{"accepted":true}

然后我们可以从OSS控制台看到备份的结果：

现在我们再往sales索引里面写一些数据：

[root@master ~]# curl -X GET "localhost:9200/_cat/indices?v"
health status index uuid pri rep docs.count docs.deleted store.size pri.store.size
green open sales 89ouBy6RQsuT34QRbn_jeQ 10 0 281502 0 15.6mb 15.6mb
green open customer fQCMEvXsQOu0UgMm1SAJlA 5 0 10000 0 717kb 717kb

我们利用刚才备份到OSS的快照来恢复sales索引，分别执行下面的命令：

[root@master ~]# curl -XPOST localhost:9200/sales/_close
{"acknowledged":true,"shards_acknowledged":true,"indices":{"sales":{"closed":true}}}
[root@master ~]# curl -XPOST 'http://localhost:9200/_snapshot/test/sales/_restore?pretty'
{
 "accepted" : true
}
[root@master ~]# curl -X GET "localhost:9200/_cat/indices?v"
health status index uuid pri rep docs.count docs.deleted store.size pri.store.size
green open sales 89ouBy6RQsuT34QRbn_jeQ 10 0 271786 0 15mb 15mb
green open customer fQCMEvXsQOu0UgMm1SAJlA 5 0 10000 0 717kb 717kb

我们可以看到，sales索引跟之前的一致。

参考资料

https://www.elastic.co/guide/en/elasticsearch/plugins/7.2/repository-s3.html

https://www.elastic.co/cn/products/elasticsearch

作者：冷月_wjh