Elasticsearch提示low disk watermark [85%] exceeded on [UTyrLH40Q9uIzHzX-yMFXg][Sonofelice][/Users/baid…(Elasticsearch prompt low disk watermark [85%] exceeded on [utyrlh40q9uizhzx ymfxg] [sonofelice] [/ users / bad)

mac本地启动es之后发现运行一段时间一分钟就能打印好几条info日志:

[2018-03-13T10:15:42,497][INFO ][o.e.c.r.a.DiskThresholdMonitor] [Sonofelice] low disk watermark [85%] exceeded on [UTyrLH40Q9uIzHzX-yMFXg][Sonofelice][/Users/baidu/Documents/work/soft/data/nodes/0] free: 15.2gb[13.4%], replicas will not be assigned to this node
[2018-03-13T10:16:12,505][INFO ][o.e.c.r.a.DiskThresholdMonitor] [Sonofelice] low disk watermark [85%] exceeded on [UTyrLH40Q9uIzHzX-yMFXg][Sonofelice][/Users/baidu/Documents/work/soft/data/nodes/0] free: 15.2gb[13.4%], replicas will not be assigned to this node

看日志的意思就是,节点sonofelice上的磁盘空间使用率超过了85%,空闲15.2gb,占总磁盘空间的13.4%。将不会在该节点上分配副本了。

虽然不影响es的正常功能使用,但是打印一堆日志够烦的,看一下是什么原因呢,能不能关掉。

第一种方式,是简单粗暴的在elasticsearch.yml文件中直接加入下面一行命令:

cluster.routing.allocation.disk.threshold_enabled: false
第二种方式,则是自己控制磁盘空间使用率超过多少的时候不再分配副本。默认是85%,所以才会有上面的日志

cluster.routing.allocation.disk.threshold_enabled: true
cluster.routing.allocation.disk.watermark.low: 30gb
cluster.routing.allocation.disk.watermark.high: 20gb

————————

After the MAC starts es locally, it is found that several info logs can be printed in a minute after running for a period of time:

[2018-03-13T10:15:42,497][INFO ][o.e.c.r.a.DiskThresholdMonitor] [Sonofelice] low disk watermark [85%] exceeded on [UTyrLH40Q9uIzHzX-yMFXg][Sonofelice][/Users/baidu/Documents/work/soft/data/nodes/0] free: 15.2gb[13.4%], replicas will not be assigned to this node
[2018-03-13T10:16:12,505][INFO ][o.e.c.r.a.DiskThresholdMonitor] [Sonofelice] low disk watermark [85%] exceeded on [UTyrLH40Q9uIzHzX-yMFXg][Sonofelice][/Users/baidu/Documents/work/soft/data/nodes/0] free: 15.2gb[13.4%], replicas will not be assigned to this node

Looking at the log means that the utilization rate of disk space on node sonofelice exceeds 85%, and the idle space is 15.2gb, accounting for 13.4% of the total disk space. The replica will not be assigned on this node.

Although it does not affect the normal function of ES, printing a pile of logs is annoying enough. Let’s see why and whether it can be turned off.

The first way is simple and rough in elastic search Add the following command line directly into the YML file:

cluster. routing. allocation. disk. threshold_ enabled: false
The second way is to control how much disk space utilization is exceeded, and no replica will be allocated. The default is 85%, which is why there are the above logs

cluster.routing.allocation.disk.threshold_enabled: true
cluster.routing.allocation.disk.watermark.low: 30gb
cluster.routing.allocation.disk.watermark.high: 20gb