golang_micro: A TSQL repository from YouDail

hackathon

A small go microservice project

run etcd3

mkdir /data/etcdV3
chmod a+rw -R /data/etcdV3

docker run -d --restart=always  -p 22379:2379 -p 22380:2380 -d -v /data/etcdV3:/data quay.io/coreos/etcd:v3

run redis


docker run --name myredis --network=host -d redis

run mysql


mkdir /usr/mysqldata
chmod a+w -R /usr/mysqldata
docker run --name mymysql --network=host -v /usr/mysqldata:/var/lib/mysql -e MYSQL_ROOT_PASSWORD=123456 -d mysql:5.7

导入数据
mysql -uroot -p123456 -h 127.0.0.1 -p 3306
source hackathon_class.sql;
source hackathon_grade.sql;
source hackathon_student.sql;

run controller & service


cd hackathon-controller
make docker
docker run -d --restart=always  --network=host hackathon-controller


cd hackathon-service
make docker
docker run -d --restart=always  --network=host hackathon-service

run gateway



cd hackathon-gateway
make docker
docker run -d --restart=always -m 4096M --memory-swap=4096M   --network=host hackathon-gateway

try it


GET http://127.0.0.1:10010/hackathon/maxClasses/2

benchmark


cd hackathon-gateway/test
go test -v -bench=".*"

使用go-wrk进行压测

执行如下命令，进行30s 1W次高并发场景模拟：

for i in {0..11}; do go-wrk -d 30 -n 10000 http://10.52.26.3:10010/hackathon/maxClasses/$i;done

使用vegeta进行压测

wget https://github.com/tsenart/vegeta/releases/download/cli%2Fv12.5.1/vegeta-12.5.1-linux-amd64.tar.gz
tar zxf vegeta-12.5.1-linux-amd64.tar.gz
mv vegeta /usr/local/bin/

先来个1600个
for m in {1..100};do for i in {0..15};do echo -e "{\"method\":\"GET\",\"url\":\"http://10.52.26.3:10010/hackathon/maxClasses/$i\"}";done;done |  vegeta  -profile heap attack -lazy -timeout 5s -workers 100 -format=json | tee results.bin | vegeta report  -type="hist[0,10ms,20ms,30ms]"
再来个1.6万个
for m in {1..1000};do for i in {0..15};do echo -e "{\"method\":\"GET\",\"url\":\"http://10.52.26.3:10010/hackathon/maxClasses/$i\"}";done;done |  vegeta  -profile heap attack -lazy -timeout 5s -workers 100 -format=json | tee results.bin | vegeta report  -type="hist[0,5ms,10ms,15ms,20ms,30ms]"
再压个16万个
for m in {1..1000};do for i in {0..15};do echo -e "{\"method\":\"GET\",\"url\":\"http://10.52.26.3:10010/hackathon/maxClasses/$i\"}";done;done |  vegeta  -profile heap attack -lazy -timeout 2s -workers 1000 -format=json | tee results.bin | vegeta report  -type=json

Go tool pprof & dstat


centos7安装graphviz包
yum install http://rpmfind.net/linux/centos/7.6.1810/os/x86_64/Packages/graphviz-2.30.1-21.el7.x86_64.rpm

查看gateway的runtime CPU和内存性能
while true;do go tool pprof --text  http://10.52.26.3:10909/debug/pprof/profile;go tool pprof --text  http://10.52.26.3:10909/debug/pprof/heap;sleep 5;done


查看系统进程级别的内存、cpu、io情况
yum install dstat -y
dstat --top-cpu --top-io --top-mem

使用go-torch生成火焰图


git clone https://github.com/brendangregg/FlameGraph.git
cp flamegraph.pl /usr/local/bin
go get -v github.com/uber/go-torch

在上面的压测过程中，使用go-torch来生成采样报告，采集时长是2分钟:

go-torch -u http://10.52.26.3:10110 -t 120

采样完成后，输出信息：Writing svg to torch.svg 表示采样成功，下面开启web服务，通过浏览器打开torch.svg查看火焰图:

python -m SimpleHTTPServer 20001

Prometheus metrics


curl http://127.0.0.1:10909/metrics

run consul：

docker run -d -p 8500:8500 consul

prometheus配置文件

# cat prometheus.yml
global:
  scrape_interval: 5s
  scrape_timeout: 5s
  evaluation_interval: 15s
scrape_configs:
  - job_name: hackathon
    relabel_configs:
    - source_labels:  ["__meta_consul_service"]
      regex: "(.*)"
      replacement: $1
      action: replace
      target_label: ""
    - source_labels:  ["__meta_consul_tags"]
      regex: "(.*),(.*),(.*)"
      replacement: $2
      action: replace
      target_label: "hackathon"
    metrics_path: /metrics
    scheme: http
    consul_sd_configs:
      - server: 10.52.26.3:8500
        scheme: http

run prometheus & granafa


docker run -d -v /usr/local/prometheus_my/prometheus.yml:/etc/prometheus/prometheus.yml -p 19090:9090   prom/prometheus

docker run -d -v /usr/local/grafana_plugins/:/var/lib/grafana/plugins -p 3000:3000 grafana/grafana

dashboards推荐240

something perfect

Do you seen some sql-script in my project ?


//从mysql5.7以后，默认的sql_mode不允许对别名字段进行group by，这里需要去掉ONLY_FULL_GROUP_BY的mode设置
//下面这行设置仅对新款库生效

set @@sql_mode='STRICT_TRANS_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION';
//下面这行设置对现有库生效
set sql_mode ='STRICT_TRANS_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION';

//统计年级id是2的所有班级的各自的男女生总人数
select classId,  sum(case when gender=1 then 1 else 0 end)  as maleCount, sum(case when gender=2 then 1 else 0 end) as female from  hackathon_student WHERE classId = 2;

//统计年级id是2的所有班级的男生人数最多的班级
select * from (select classId,  sum(case when gender=1 then 1 else 0 end)  as male from  hackathon_student WHERE classId in ( SELECT id FROM hackathon_class WHERE gradeId = 2 ) group by classId) a order by male desc limit 1;

//统计年级id是2的所有班级的女生人数最多的班级
select * from (select classId,  sum(case when gender=2 then 1 else 0 end) as female from  hackathon_student WHERE classId in ( SELECT id FROM hackathon_class WHERE gradeId = 2 ) group by classId) a order by female desc limit 1;