Abnormal index_writer_memory,its so high!

Elastic Stack Elasticsearch
1

Our ElasticSearch version is v1.6.2.
31G jvm memory in 156G machine total memory.
Recently I found out our node is OOM frequently,then I found an index use "index_writer_memory" is so high.

截屏2020-03-0923.51.06
截屏2020-03-0923.51.062482×448 27.8 KB

It really confuse me,because this index is not a bigdata index,look below.
截屏2020-03-0923.55.08
截屏2020-03-0923.55.081736×448 15.7 KB

and at the memory flood out time,the index rate is not high at all.
截屏2020-03-0923.57.53
截屏2020-03-0923.57.531736×448 18.6 KB

Can anybody tell me whats going on and how can I fix this problem?

2

This version reached the end of its life many years ago, and more recent versions have seen big improvements in resiliency. The very first thing you should do is upgrade.

3

I know the version is out of times,but this problem is in our production environment so I have to fix it first.