| Server IP : 112.74.42.100 / Your IP : 216.73.216.142 Web Server : nginx/1.20.2 System : Linux iZwz96lx4pjqiy84w4hni7Z 5.10.134-17.3.al8.x86_64 #1 SMP Thu Oct 31 14:29:57 CST 2024 x86_64 User : www ( 1000) PHP Version : 8.0.26 Disable Function : passthru,exec,system,putenv,chroot,chgrp,chown,shell_exec,popen,proc_open,pcntl_exec,ini_alter,ini_restore,dl,openlog,syslog,readlink,symlink,popepassthru,pcntl_alarm,pcntl_fork,pcntl_waitpid,pcntl_wait,pcntl_wifexited,pcntl_wifstopped,pcntl_wifsignaled,pcntl_wifcontinued,pcntl_wexitstatus,pcntl_wtermsig,pcntl_wstopsig,pcntl_signal,pcntl_signal_dispatch,pcntl_get_last_error,pcntl_strerror,pcntl_sigprocmask,pcntl_sigwaitinfo,pcntl_sigtimedwait,pcntl_exec,pcntl_getpriority,pcntl_setpriority,imap_open,apache_setenv MySQL : OFF | cURL : ON | WGET : ON | Perl : ON | Python : ON | Sudo : ON | Pkexec : ON Directory : /usr/share/doc/keentuned/ |
Upload File : |
# QuickStart ###### Version: 2.3.0 ## Dependencies 安装python3运行环境 ```s yum install python36 python36-devel pip3 install --upgrade pip ``` 安装python依赖包 ```s pip3 install hyperopt==0.1.2 pip3 install numpy==1.19.5 pip3 install POAP==0.1.26 pip3 install pySOT==0.3.3 pip3 install scikit_learn==1.1.1 pip3 install setuptools==39.2.0 pip3 install shap==0.40.0 pip3 install tornado==6.1 pip3 install xgboost==1.5.2 pip3 install pynginxconfig==0.3.4 pip3 install torch ``` ## Installation ### 配置yum源 将以下keentune源追加进系统的yum源 ```s vim /etc/yum.repos.d/epel.repo ``` ```s [keentune] name=keentune-os baseurl=https://mirrors.openanolis.cn/anolis/8.4/Plus/x86_64/os/ gpgkey=https://mirrors.openanolis.cn/anolis/RPM-GPG-KEY-ANOLIS enabled=1 ``` 更新yum源缓存 ```s yum clean all yum makecache ``` ### 安装KeenTune 使用yum安装KeenTune各组件，使用源码安装的方法参考[这里](./2.Installation_cn.md) ``` yum install keentuned keentune-brain keentune-bench keentune-target ``` ## Usages ### 启动KeenTune ```s # 启动keentuned服务 systemctl start keentuned # 启动keentune-brain服务 systemctl start keentune-brain # 启动keentune-bench服务 systemctl start keentune-bench # 启动keentune-target服务 systemctl start keentune-target ``` ### Auto-Tuning(参数调优) 使用[默认配置](./install/Configuration_cn.md)开始一次[参数调优](./how-to/How%20to%20Auto-Tuning%20Knobs_cn.md)，总轮次为10轮。 ```s keentune param tune --job tune_demo --iteration 10 ``` ### Profile(专家调优方案) 为[默认配置](./install/Configuration_cn.md)中[Group1](./design/target_group_cn.md)的机器设置[CPU高负载](./design/buildin_profile_cn.md)的[专家调优](./how-to/How%20to%20Auto-Tuning%20Knobs_cn.md)方案。 ```s keentune profile set --group1 cpu_high_load.conf ``` ### Sensitize Knobs(敏感参数识别) 使用[默认配置](./install/Configuration_cn.md)和Auto-Tuning中产生的数据对[参数敏感性](./how-to/How%20to%20sentivize%20knob.md)进行识别 ```s keentune sensitize train --data tune_demo --output tune_demo ``` ### Pressure Control(智能控压) 在http长链接工作负载下对[wrk](./5.benchmkar_wrk.md)的参数进行[智能化控制](./how-to/How%20to%20sentivize%20knob.md)，总轮次为10轮。 ```s keentune param tune --param wrk.json --bench wrk_nginx_long.json --job wrk_demo --iteration 10 ```