root用户编辑
vi /etc/sysctl.conf
vm.max_map_count=262144
vi /etc/security/limits.conf
admin hard nofile 65536
admin soft nofile 65536
admin soft memlock unlimited
admin hard memlock unlimited
admin soft nproc 65536
admin hard nproc 65536
sysctl -p 立即生效
admin用户重新登录：
ulimit -a 查看确认
1
memory locking requested for elasticsearch process but memory is not locked
配置了bootstrap.memory_lock:true
https://www.elastic.co/guide/en/elasticsearch/reference/current/_memory_lock_check.html
max number of threads [1024] for user [admin] is too low, increase to at least [2048]
ulimit -u 2048
nproc –>/etc/security/limits.conf
https://www.elastic.co/guide/en/elasticsearch/reference/master/max-number-of-threads.html
[WARN ][o.e.b.JNANatives ] unable to install syscall filter:
java.lang.UnsupportedOperationException: seccomp unavailable: requires kernel 3.5+ with CONFIG_SECCOMP and CONFIG_SECCOMP_FILTER compiled in
max virtual memory areas vm.max_map_count [65530] is too low, increase to at least [262144]
sudo sysctl -w vm.max_map_count=262144
/etc/sysctl.conf
vm.max_map_count=262144
max file descriptors [4096] for elasticsearch process is too low, increase to at least [65536]
临时解决： sudo ulimit -n 65536
若找不到ulimit命令(部分Ubuntu系统
sudo sh -c "ulimit -n 65536 && exec su $LOGNAME"
vi /etc/security/limits.conf
admin hard nofile 65536
admin soft nofile 65536
改为还ulimint -n 检查还是没有生效，特别是Ubuntu，参考
常见
异常#
注意：yml配置文件的值前有一个空格隔开
max number of threads [1024] for user [admin] is too low, increase to at least [2048]
原因：用户允许最大线程数首先
vi bin/elasticsearch
ulimit -u 2048 //仅供测试使用
Warning: Ignoring JAVA_OPTS=…….
Please pass JVM parameters via ES_JAVA_OPTS instead
vi bin/elasticsearch
添加如下配置项：
JAVA_HOME=”/export/servers/jdk1.8.0_60”
JAVA_OPTS=”“
JAVA_OPTS配置为空，是为了不受系统配置的环境变量的影响
can not run elasticsearch as root
不能以root用户启动ES服务器。非要以root用户运行？
配置 -Des.insecure.allow.root=true
对于5.X，在config/jvm.options配置文件中，添加-Des.insecure.allow.root=true
内存锁定:
解决方法1：配置 config/elasticsearch.yml ，注释掉以下内容
#bootstrap.memory_lock: true
解决方法2：配置：/etc/security/limits.conf,
admin soft memlock unlimited
admin hard memlock unlimited
system call filters failed to install; check the logs and fix your configuration or disable system call filters at your own risk
Your kernel does not support seccomp.
Elasticsearch attempts to utilize seccomp by default (via the setting bootstrap.system_call_filter).
Starting in 5.2.0, if you’re in production mode, bootstrap.system_call_filter is enabled, and initializing seccomp fails, then Elasticsearch will refuse to bootstrap.
You either have to migrate to a kernel that supports seccomp, or disable bootstrap.system_call_filter.
Centos6不支持SecComp，而ES5.2.0默认bootstrap.system_call_filter为true
禁用：在elasticsearch.yml中配置bootstrap.system_call_filter为false，注意要在Memory下面:
bootstrap.memory_lock: false
bootstrap.system_call_filter: false
配置IP和http端口，TCP端口默认在HTTP端口上加100
network.host: 192.168.179.20
http.port: 9201
path.conf is not a recognized option
之前的配置：
--path.conf=${ES_CONF}
修改为：
-Epath.conf=${ES_CONF}
unknown setting [path.plugins]
https://www.elastic.co/guide/en/elasticsearch/reference/5.0/breaking_50_plugins.html#_custom_plugins_path
specify a custom plugins path via path.plugins has been removed.
node settings must not contain any index level settings
Since elasticsearch 5.x index level settings can NOT be set on the nodes
configuration like the elasticsearch.yaml, in system properties or command line
arguments.
curl -XPUT ‘http://localhost:9200/_all/_settings?preserve_existing=true’ -d ‘{
“index.number_of_shards” : “3”
}’
curl -XPUT ‘http://localhost:9200/_all/_settings?preserve_existing=true’ -d ‘{
“index.mapper.dynamic” : “false”,
“index.translog.durability” : “async”,
“index.translog.sync_interval” : “30s”
}’
即以index开头的配置删除。
unknown setting [bootstrap.mlockall]
修改为： bootstrap.memory_lock: true
unknown setting [action.disable_delete_all_indices]
新配置：action.destructive_requires_name: true
