我无法弄清楚为什么会发生这种情况以及如何解决这个问题:
请注意,分配了所有分片的第一个logstash索引仅在该状态,因为我手动分配了它们。
直到今天,我发现所有的logstash索引都没有分配碎片,所以一切都按预期工作了几个月。
使用Elasticsearch v1.4.0
我的Elasticsearch日志在2015年7月29日之前显示如下:
[2015-07-29 00:01:53,352][DEBUG][discovery.zen.publish ] [Thor] received cluster state version 4827
此外,服务器上运行的脚本会将logstash索引的数量减少到30个。我认为这是此类日志行的来源:
[2015-07-29 17:43:12,800][DEBUG][gateway.local.state.meta ] [Thor] [logstash-2015.01.25] deleting index that is no longer part of the metadata (indices: [[kibana-int, logstash-2015.07.11, logstash-2015.07.04, logstash-2015.07.03, logstash-2015.07.12, logstash-2015.07.21, logstash-2015.07.17, users_1416418645233, logstash-2015.07.20, logstash-2015.07.25, logstash-2015.07.06, logstash-2015.07.28, logstash-2015.01.24, logstash-2015.07.18, logstash-2015.07.26, logstash-2015.07.08, logstash-2015.07.19, logstash-2015.07.09, logstash-2015.07.22, logstash-2015.07.07, logstash-2015.07.29, logstash-2015.07.10, logstash-2015.07.05, logstash-2015.07.01, logstash-2015.07.16, logstash-2015.07.24, logstash-2015.07.02, logstash-2015.07.27, logstash-2015.07.14, logstash-2015.07.13, logstash-2015.07.23, logstash-2015.06.30, logstash-2015.07.15]])
7月29日,有一些我以前见过的新作品:
[2015-07-29 00:01:38,024][DEBUG][action.bulk ] [Thor] observer: timeout notification from cluster service. timeout setting [1m], time since start [1m]
[2015-07-29 17:12:58,658][INFO ][cluster.routing.allocation.decider] [Thor] updating [cluster.routing.allocation.enable] from [ALL] to [NONE]
[2015-07-29 17:12:58,658][INFO ][cluster.routing.allocation.decider] [Thor] updating [cluster.routing.allocation.disable_allocation] from [false] to [true]