关于FATE联邦框架部署FATE-Serving在线推理中遇到zookeeper问题的解决办法

首先看两张出现错误的图片

我也是弄了好久 最后用以下办法解决了。

一、查看相关配置

具体的话就是检查Fate-serving在线推理的部署设置，网上很多部署案例。想知道可以看我以后发的文章。

二、具体操作

进入虚拟环境

source /data/projects/fate/bin/init_env.sh
cd /data/projects/fate

fate-serving启动zookeeper

cd /data/projects/apache-zookeeper-3.6.2-bin/bin/
./zkServer.sh restart

或者

root@localhost bin]# sh ./zkServer.sh restart
/usr/bin/java
ZooKeeper JMX enabled by default
Using config: /data/projects/apache-zookeeper-3.6.2-bin/bin/../conf/zoo.cfg
/usr/bin/java
ZooKeeper JMX enabled by default
Using config: /data/projects/apache-zookeeper-3.6.2-bin/bin/../conf/zoo.cfg
Stopping zookeeper ... STOPPED
/usr/bin/java
ZooKeeper JMX enabled by default
Using config: /data/projects/apache-zookeeper-3.6.2-bin/bin/../conf/zoo.cfg
Starting zookeeper ... STARTED
[root@localhost bin]# 

1. 开启各种配置

cd /data/projects/common/supervisord

启动/关闭/重启/查看所有：
#注意：因mysql是基础组件，启动较慢，建议重启操作是先停止所有组件，然后先启动mysql，再启动其他组件
sh service.sh /stop/restart/status all 

sh service.sh restart all

#说明：因为fateflow依赖的组件比较多，重启所有的操作可能会导致fateflow启动异常，处理如下：
netstat -tlnp | grep 9360
如果没有端口则重起fateflow：
sh service.sh stop fate-fateflow
sh service.sh start fate-fateflow

启动/关闭/重启/查看单个模块(可选：clustermanager，nodemanager，rollsite，fateflow，fateboard，mysql)：

sh service.sh start/stop/restart/status fate-clustermanager

2. fate-serving启动server

cd /data/projects/fate-serving/server
sh service.sh restart

[root@localhost bin]# cd /data/projects/fate-serving/server
[root@localhost server]# sh service.sh restart
killing:    PID TTY          TIME CMD
 60010 pts/10   00:09:40 java
please wait
killed
try to start serving-server
service start sucessfully. pid: 67992
status: UID         PID   PPID  C STIME TTY          TIME CMD
root      67992  67951 99 19:12 pts/11   00:00:08 java -Dspring.config.location=/data/projects/fate-serving/server/conf/serving-server.properties -cp conf/:lib/*:extension/*:fate-serving-server.jar com.webank.ai.fate.serving.Bootstrap

3.fate-serving启动proxy

cd /data/projects/fate-serving/proxy/
sh service.sh restart

[root@localhost server]# cd /data/projects/fate-serving/proxy/
[root@localhost proxy]# sh service.sh restart
killing:    PID TTY          TIME CMD
 60863 pts/10   00:09:00 java
please wait
killed
try to start serving-proxy
service start sucessfully. pid: 68145
status: UID         PID   PPID  C STIME TTY          TIME CMD
root      68145  68106 99 19:13 pts/11   00:00:11 java -Dspring.config.location=/data/projects/fate-serving/proxy/conf/application.properties -cp conf/:lib/*:fate-serving-proxy.jar com.webank.ai.fate.serving.proxy.bootstrap.Bootstrap

4.fate-serving启动admin

cd /data/projects/fate-serving/admin/
sh service.sh restart

[root@localhost proxy]# cd /data/projects/fate-serving/admin/
[root@localhost admin]# sh service.sh restart
killing:    PID TTY          TIME CMD
 61065 pts/10   00:09:19 java
please wait
killed
try to start serving-admin
service start sucessfully. pid: 68294
status: UID         PID   PPID  C STIME TTY          TIME CMD
root      68294  68257 99 19:13 pts/11   00:00:12 java -Dspring.config.location=/data/projects/fate-serving/admin/conf/application.properties -cp conf/:lib/*:fate-serving-admin.jar com.webank.ai.fate.serving.admin.Bootstrap

5.开启fate-flow

cd /data/projects/fate/python/fate_flow
sh service.sh start fate-fateflow

[root@localhost fate_flow]# sh service.sh start fate-fateflow
PROJECT_BASE: /data/projects/fate
PYTHONPATH: /data/projects/fate/python:/data/projects/fate/eggroll/python
EGGROLL_HOME: /data/projects/fate/eggroll
SPARK_HOME: 
found service conf: /data/projects/fate/conf/service_conf.yaml
fate flow http port: 9380, grpc port: 9360

service already started. pid: 2650
status:app        2650  2.3  1.0 3151468 87728 ?       Sl   3月15 231:35 python /data/projects/fate/python/fate_flow/fate_flow_server.py
python   2650  app   16u  IPv4   36440      0t0  TCP localhost.localdomain:boxp (LISTEN)
python  2650  app   13u  IPv6    36439      0t0  TCP localhost.localdomain:9360 (LISTEN)
[root@localhost fate_flow]# 

解决