多个zookeeper节点导致内存不足问题？
[英]Multiple zookeeper nodes causing out of memory issue?
问题描述
I am running 3 zookeeper instances on two servers as a cluster of 6 and i am running zookeeper using zkServer.sh script without any Java parameter on 64 bit Red Hat Linux with 12G of RAM on each server. 
我在两个服务器上运行3个zookeeper实例作为6的集群，我使用zkServer.sh脚本运行zookeeper，在每个服务器上使用12位RAM的64位Red Hat Linux上没有任何Java参数。
Now as i understood that for each unspecified java process, it takes around 25% of total RAM and i have 3 zookeeper instances, i could see that after a sometime, i get connection reset errors with zookeepers and very low memory left. 现在，据我了解，对于每个未指定的java进程，它占用总RAM的大约25％，并且我有3个zookeeper实例，我可以看到在一段时间之后，我得到动物园管理员的连接重置错误和非常低的内存。
Can anybody suggest how can i control zookeeper nodes limited by predefined memory specs such as -Xms & -Xmx to avoid these issues as my cluster does not work properly due to this issue? 任何人都可以建议我如何控制受预定义内存规范（如-Xms和-Xmx）限制的zookeeper节点以避免这些问题，因为我的集群由于此问题而无法正常工作？
Regards. 问候。
1 个解决方案
解决方案1
3 已采纳 2014-09-16 11:23:12
要向zookeeper添加JVM参数，请在每个具有以下内容的zookeeper中的conf文件夹下创建名为java.env的文件。
export JVMFLAGS="-Xms512m -Xmx512m"
声明:本站的技术帖子网页，遵循CC BY-SA 4.0协议，如果您需要转载，请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.