Elasticsearch6.13 升级6.24 单节点停机升级

Elasticsearch6.x 升级6.y 是支持滚动升级的，目前我们测试环境只有一个节点只能停机升级了

准备工作

禁用分片分配

curl -X PUT "localhost:9200/_cluster/settings" -H 'Content-Type: application/json' -d'
{
"persistent": {
"cluster.routing.allocation.enable": "none"
}
}
'

执行同步刷新，加速碎片恢复

curl -X POST "localhost:9200/_flush/synced"

关闭老版本

supervisorctl stop elasticsearch

启动新版本

下载安装包

wget https://artifacts.elastic.co/downloads/elasticsearch/elasticsearch-6.2.4.tar.gz
tar -zvxf elasticsearch-6.2..tar.gz -C /data/app/
chown -R elasticsearch.elasticsearch elasticsearch-6.2.4

拷贝配置文件和数据目录

cp -rf elasticsearch-6.1./config/ elasticsearch-6.2./
cp -rf elasticsearch-6.1./data/ elasticsearch-6.2./

修改配置文件中数据的目录

vim elasticsearch-6.2./config/elasticsearch.yml

# Path to directory where to store the data (separate multiple locations by comma):
#
 path.data: /data/app/elasticsearch/data

启动新版本

vim /etc/supervisor/conf.d/es-new.conf

[program:elasticsearch-new]
directory=/data/app/elasticsearch-6.2.4/bin/
command=su - elasticsearch -c "/data/app/elasticsearch-6.2.4/bin/elasticsearch"
autostart=true
startsecs=
autorestart=true
startretries=
user=root
priority=
redirect_stderr=true
stdout_logfile_maxbytes=50MB
stdout_logfile_backups=
stdout_logfile=/var/log/elasticsearch/es-new_supervisor.out
stopasgroup=false
killasgroup=false

supervisorctl reload
supervisorctl start elasticsearch-new
supervisorctl status

elasticsearch                    STOPPED   May 22 12:58 PM
elasticsearch-head               RUNNING   pid , uptime ::
elasticsearch-new                RUNNING   pid , uptime ::
kibana                           RUNNING   pid , uptime ::

等待集群状态健康状态为yellow就行了

curl -X GET "localhost:9200/_cat/health"

 :: elasticsearchbigdata yellow         - 50.0%

curl -X GET "localhost:9200/_cat/recovery"

product_brand_100301_shop_330778_current  70ms  existing_store done n/a n/a 192.168.101.55 elasticsearch n/a n/a   100.0%     100.0%       100.0%
shop_brand_100479_current                 44ms  existing_store done n/a n/a 192.168.101.55 elasticsearch n/a n/a   100.0%     100.0%        100.0%

#恢复分配

curl -X PUT "localhost:9200/_cluster/settings" -H 'Content-Type: application/json' -d'
{
  "persistent": {
    "cluster.routing.allocation.enable": null
  }
}
'