前言
Metersphere部署过程中遇到的问题有点多，原因是其容器的架构蛮复杂的，比较容易踩坑，所以记录一下。
介绍
MeterSphere 是开源持续测试平台，遵循 GPL v3 开源许可协议，涵盖测试管理、接口测试、UI 测试和性能测试等功能，全面兼容 JMeter、Selenium 等主流开源标准，有效助力开发和测试团队充分利用云弹性进行高度可扩展的自动化测试，加速高质量的软件交付。
主页：https://metersphere.io/
源码：https://github.com/metersphere/metersphere
前置环境
Ubuntu 20.04
Docker version 27.3.1
docker-compose version 1.27.4
openjdk version “17.0.13”
MeterSphere-v2.4.0安装部署
首先以v2.4.0为例，直接给出我的sh：
if [ "$(id -u)" -ne 0 ]; then
echo "此脚本需要root权限，请使用sudo执行。"
exit 1
fi
wget https://github.com/metersphere/metersphere/releases/download/v2.4.0/metersphere-online-installer-v2.4.0.tar.gz --no-check-certificate
tar zxvf metersphere-online-installer-v2.4.0.tar.gz
cd metersphere-online-installer-v2.4.0
#启用工作台
#CONFIG_FILE="install.conf"
#echo "修改配置文件..."
#sed -i 's/^MS_UI_ENABLED=false$/MS_UI_ENABLED=true/' "$CONFIG_FILE"
#sed -i 's/^MS_WORKSTATION_ENABLED=false$/MS_WORKSTATION_ENABLED=true/' "$CONFIG_FILE"
#MYSQL降级
#sed -i 's/\${MS_IMAGE_PREFIX}\/mysql:8.0.33/\${MS_IMAGE_PREFIX}\/mysql:8.0.30/' "metersphere/docker-compose-mysql.yml"
sudo ./install.sh
msctl status
可以使用命令msctl status检查服务运行情况，正确部署后的输出如下所示：
Name Command State Ports
------------------------------------------------------------------------------------------------------
api-test /deployments/run-java.sh Up (healthy) 0.0.0.0:10000->10000/tcp,:::10000
->10000/tcp, 0.0.0.0:10001->10001
/tcp,:::10001->10001/tcp, 0.0.0.0
:10002->10002/tcp,:::10002->10002
/tcp, 0.0.0.0:10003->10003/tcp,::
:10003->10003/tcp, 0.0.0.0:10004-
>10004/tcp,:::10004->10004/tcp, 0
.0.0.0:10005->10005/tcp,:::10005-
>10005/tcp, 0.0.0.0:10006->10006/
tcp,:::10006->10006/tcp, 0.0.0.0:
10007->10007/tcp,:::10007->10007/
tcp, 0.0.0.0:10008->10008/tcp,:::
10008->10008/tcp, 0.0.0.0:10009->
10009/tcp,:::10009->10009/tcp, 0.
0.0.0:10010->10010/tcp,:::10010->
10010/tcp
eureka /deployments/run-java.sh Up (healthy)
gateway /deployments/run-java.sh Up (healthy) 0.0.0.0:8081->8000/tcp,:::8081->8
000/tcp
kafka /opt/bitnami/scripts/kafka ... Up (healthy) 0.0.0.0:9092->9092/tcp,:::9092->9
092/tcp
minio /usr/bin/docker-entrypoint ... Up (healthy) 0.0.0.0:9000->9000/tcp,:::9000->9
000/tcp, 0.0.0.0:9001->9001/tcp,:
::9001->9001/tcp
ms-data-streaming /deployments/run-java.sh Up (healthy)
ms-node-controller /deployments/run-java.sh Up (healthy) 0.0.0.0:8082->8082/tcp,:::8082->8
082/tcp, 0.0.0.0:9100->9100/tcp,:
::9100->9100/tcp
ms-prometheus /bin/prometheus --config.f ... Up (healthy) 0.0.0.0:9090->9090/tcp,:::9090->9
090/tcp
mysql docker-entrypoint.sh mysqld Up (healthy) 0.0.0.0:3307->3306/tcp,:::3307->3
306/tcp, 33060/tcp
nodeexporter /bin/node_exporter --path. ... Up (healthy)
performance-test /deployments/run-java.sh Up (healthy)
project-management /deployments/run-java.sh Up (healthy)
redis docker-entrypoint.sh redis ... Up (healthy) 0.0.0.0:6379->6379/tcp,:::6379->6
379/tcp
report-stat /deployments/run-java.sh Up (healthy)
system-setting /deployments/run-java.sh Up (healthy)
test-track /deployments/run-java.sh Up (healthy)
URL: http://$LOCAL_IP:8081
用户名: admin
初始密码: metersphere
然而整个过程并非这么顺利的，过程中会出现各种各样的问题（尤其是在服务器部署的情况下），我的建议是时刻检查安装目录在哪里，通过Debug找到到底是哪里的conf文件在生效，免得乱改一通发现压根没从这里读配置。。。
接下来说一些比较好解决的问题：
v2.4.0 Q&A
Q：部分容器（例如api-test等）起不来，msctl status显示mysql容器一直在restarting
$ msctl status
Name Command State Ports
------------------------------------------------------------------------------------------------------------------------------------------
eureka /deployments/run-java.sh Up (healthy)
gateway /deployments/run-java.sh Up (healthy) 0.0.0.0:8081->8000/tcp,:::8081->8000/tcp
kafka /opt/bitnami/scripts/kafka ... Up (healthy) 0.0.0.0:9092->9092/tcp,:::9092->9092/tcp
minio /usr/bin/docker-entrypoint ... Up (healthy) 0.0.0.0:9000->9000/tcp,:::9000->9000/tcp,
0.0.0.0:9001->9001/tcp,:::9001->9001/tcp
ms-data-streaming /deployments/run-java.sh Up (healthy)
ms-node-controller /deployments/run-java.sh Up (healthy) 0.0.0.0:8082->8082/tcp,:::8082->8082/tcp,
0.0.0.0:9100->9100/tcp,:::9100->9100/tcp
ms-prometheus /bin/prometheus --config.f ... Up (healthy) 0.0.0.0:9090->9090/tcp,:::9090->9090/tcp
mysql docker-entrypoint.sh mysqld Restarting
nodeexporter /bin/node_exporter --path. ... Up (healthy)
redis docker-entrypoint.sh redis ... Up (healthy) 0.0.0.0:6379->6379/tcp,:::6379->6379/tcp
A：这个问题一般出现在之前安装过高版本Metersphere的情况，看一下日志
$ docker logs mysql
2025-01-09 04:58:20+00:00 [Note] [Entrypoint]: Entrypoint script for MySQL Server 8.0.30-1.el8 started.
2025-01-09 04:58:20+00:00 [Note] [Entrypoint]: Switching to dedicated user 'mysql'
2025-01-09 04:58:20+00:00 [Note] [Entrypoint]: Entrypoint script for MySQL Server 8.0.30-1.el8 started.
'/var/lib/mysql/mysql.sock' -> '/var/run/mysqld/mysqld.sock'
2025-01-09T04:58:20.857320Z 0 [Warning] [MY-011068] [Server] The syntax '--skip-host-cache' is deprecated and will be removed in a future release. Please use SET GLOBAL host_cache_size=0 instead.
2025-01-09T04:58:20.857349Z 0 [Warning] [MY-011068] [Server] The syntax 'expire-logs-days' is deprecated and will be removed in a future release. Please use binlog_expire_logs_seconds instead.
2025-01-09T04:58:20.857466Z 0 [System] [MY-010116] [Server] /usr/sbin/mysqld (mysqld 8.0.30) starting as process 1
2025-01-09T04:58:20.861214Z 1 [System] [MY-013576] [InnoDB] InnoDB initialization has started.
2025-01-09T04:58:21.299938Z 1 [ERROR] [MY-013171] [InnoDB] Cannot boot server version 80030 on data directory built by version 80033. Downgrade is not supported
mysqld: Can't open file: 'mysql.ibd' (errno: 0 - )
2025-01-09T04:58:21.594261Z 1 [ERROR] [MY-010334] [Server] Failed to initialize DD Storage Engine
2025-01-09T04:58:21.594352Z 0 [ERROR] [MY-010020] [Server] Data Dictionary initialization failed.
2025-01-09T04:58:21.594359Z 0 [ERROR] [MY-010119] [Server] Aborting
2025-01-09T04:58:21.594589Z 0 [System] [MY-010910] [Server] /usr/sbin/mysqld: Shutdown complete (mysqld 8.0.30) MySQL Community Server - GPL.
可以看到是mysql版本不匹配问题，修改metersphere-online-installer-v2.x.x/metersphere/docker-compose-mysql.yml，进行降级：
version: "2.1"
services:
mysql:
image: ${MS_IMAGE_PREFIX}/mysql:8.0.33
container_name: mysql
healthcheck:
test: ["CMD", "mysqladmin", "ping", "-h", "localhost", "-uroot", "-p${MS_MYSQL_PASSWORD}"]
interval: 6s
timeout: 5s
retries: 50
environment:
MYSQL_DATABASE: ${MS_MYSQL_DB}
MYSQL_ROOT_PASSWORD: ${MS_MYSQL_PASSWORD}
ports:
- 3307:3306
volumes:
- ${MS_BASE}/metersphere/conf/my.cnf:/etc/mysql/conf.d/my.cnf
- ${MS_BASE}/metersphere/bin/mysql:/docker-entrypoint-initdb.d/
- ms-mysql-data:/var/lib/mysql
networks:
- ms-network
restart: always
system-setting:
depends_on:
mysql:
condition: service_healthy
volumes:
ms-mysql-data:
driver_opts:
type: none
device: ${MS_BASE}/metersphere/data/mysql
o: bind
修改之后重新安装即可，当然也可以直接解除start_in_docker.sh里MYSQL降级部分的注释后重新执行即可。
sed -i 's/\${MS_IMAGE_PREFIX}\/mysql:8.0.33/\${MS_IMAGE_PREFIX}\/mysql:8.0.30/' "metersphere/docker-compose-mysql.yml"
Q：工作台未启用，功能不齐全，左侧边栏只显示两种功能。
A：修改install.conf如下内容：
MS_UI_ENABLED=true
MS_WORKSTATION_ENABLED=true
执行msctl reload后，即可看到工作台容器已经启动便可以正常使用，访问UI后可以看到左侧边栏已经显示全面。当然也可以直接解除start_in_docker.sh里启用工作台部分的注释后重新执行即可。
CONFIG_FILE="install.conf"
sed -i 's/^MS_UI_ENABLED=false$/MS_UI_ENABLED=true/' "$CONFIG_FILE"
sed -i 's/^MS_WORKSTATION_ENABLED=false$/MS_WORKSTATION_ENABLED=true/' "$CONFIG_FILE"
MeterSphere-v2.10.1-lts安装部署
再拿v2.10.1-lts版本举例：
$ wget https://github.com/metersphere/metersphere/releases/download/v2.10.1-lts/metersphere-online-installer-v2.10.1-lts.tar.gz --no-check-certificate
$ tar zxvf metersphere-online-installer-v2.10.1-lts.tar.gz
$ cd metersphere-online-installer-v2.10.1-lts
$ sudo ./install.sh
$ msctl status
Name Command State Ports
--------------------------------------------------------------------------------------
api-test /deployments/run-java.sh Up (healthy) 0.0.0.0:10000->10000/tc
p,:::10000->10000/tcp,
0.0.0.0:10001->10001/tc
p,:::10001->10001/tcp,
0.0.0.0:10002->10002/tc
p,:::10002->10002/tcp,
0.0.0.0:10003->10003/tc
p,:::10003->10003/tcp,
0.0.0.0:10004->10004/tc
p,:::10004->10004/tcp,
0.0.0.0:10005->10005/tc
p,:::10005->10005/tcp,
0.0.0.0:10006->10006/tc
p,:::10006->10006/tcp,
0.0.0.0:10007->10007/tc
p,:::10007->10007/tcp,
0.0.0.0:10008->10008/tc
p,:::10008->10008/tcp,
0.0.0.0:10009->10009/tc
p,:::10009->10009/tcp,
0.0.0.0:10010->10010/tc
p,:::10010->10010/tcp
eureka /deployments/run-java.sh Up (healthy)
gateway /deployments/run-java.sh Up (healthy) 0.0.0.0:8081->8000/tcp,
:::8081->8000/tcp
kafka /opt/bitnami/scripts/kaf Up (healthy) 0.0.0.0:9092->9092/tcp,
ka ... :::9092->9092/tcp
minio /usr/bin/docker- Up (healthy) 0.0.0.0:9000->9000/tcp,
entrypoint ... :::9000->9000/tcp, 0.0.
0.0:9001->9001/tcp,:::9
001->9001/tcp
ms-data-streaming /deployments/run-java.sh Up (healthy)
ms-node-controller sh -c sed -i Up (healthy) 0.0.0.0:8082->8082/tcp,
"s/:101:/:136 ... :::8082->8082/tcp, 0.0.
0.0:9100->9100/tcp,:::9
100->9100/tcp
ms-prometheus /bin/prometheus Up (healthy) 0.0.0.0:9091->9090/tcp,
--config.f ... :::9091->9090/tcp
mysql docker-entrypoint.sh Up (healthy) 0.0.0.0:3306->3306/tcp,
mysqld :::3306->3306/tcp,
33060/tcp
nodeexporter /bin/node_exporter Up (healthy)
--path. ...
performance-test /deployments/run-java.sh Up (healthy)
project-management /deployments/run-java.sh Up (healthy)
redis docker-entrypoint.sh Up (healthy) 0.0.0.0:6379->6379/tcp,
redis ... :::6379->6379/tcp
report-stat /deployments/run-java.sh Up (healthy)
system-setting /deployments/run-java.sh Up (healthy)
test-track /deployments/run-java.sh Up (healthy)
workstation /deployments/run-java.sh Up (healthy)
v2.10.1-lts Debug
Debug1：访问Web UI有{"success":false,"message":"401 UNAUTHORIZED \"Not found session, Please Login again.\"","data":null}报错，一定要等待所有容器Up并healthy状态，后再等5min访问Web UI（不要中途切换）。
Debug2：9090端口号占用问题，在docker-compose-prometheus.yml和install.conf修改为9091即可。
# Debug3: Additionally
$ msctl restart gateway
$ msctl restart workstation
$ msctl restart prometheus