当前位置：首页 > backend >正文

TencentOS Server 4.4 下创建mysql容器无法正常运行的问题

backend 2025/8/28 11:46:05

环境

腾讯的 TencentOS Server 4.4 服务器系统

Linux app 6.6.92-34.1.tl4.x86_64 #1 SMP PREEMPT_DYNAMIC Wed Jun 25 14:33:47 CST 2025 x86_64 x86_64 x86_64 GNU/Linux

docker使用的是yum安装的版本

[root@app ~]# docker version
Client:Version:           28.0.1-20241223130549-3b49debAPI version:       1.48Go version:        go1.24.4Git commit:        3b49debBuilt:             Mon Aug 11 09:03:23 2025OS/Arch:           linux/amd64Context:           defaultServer:Engine:Version:          28.0.1API version:      1.48 (minimum version 1.24)Go version:       go1.24.2Git commit:       %{_gitcommit_engine}Built:            Thu Jun 12 10:00:10 2025OS/Arch:          linux/amd64Experimental:     falsecontainerd:Version:          1.7.27GitCommit:        runc:Version:          1.1.14GitCommit:        docker-init:Version:          0.19.0GitCommit:        de40ad0

问题现象

docker创建的mysql 5.7.44容器（官方镜像）无法正常启动，隔几秒就重启

docker logs mysql 打印：

2025-08-27 15:45:46+08:00 [Note] [Entrypoint]: Entrypoint script for MySQL Server 5.7.44-1.el7 started.
2025-08-27 15:46:01+08:00 [ERROR] [Entrypoint]: mysqld failed while attempting to check configcommand was: mysqld --verbose --help --log-bin-index=/tmp/tmp.RBl3Hbmt6u2025-08-27 15:46:15+08:00 [Note] [Entrypoint]: Entrypoint script for MySQL Server 5.7.44-1.el7 started.
2025-08-27 15:46:26+08:00 [ERROR] [Entrypoint]: mysqld failed while attempting to check configcommand was: mysqld --verbose --help --log-bin-index=/tmp/tmp.hZa5kwNZnP2025-08-27 15:46:28+08:00 [Note] [Entrypoint]: Entrypoint script for MySQL Server 5.7.44-1.el7 started.
2025-08-27 15:46:37+08:00 [ERROR] [Entrypoint]: mysqld failed while attempting to check configcommand was: mysqld --verbose --help --log-bin-index=/tmp/tmp.6eeHXfFfzH2025-08-27 15:46:39+08:00 [Note] [Entrypoint]: Entrypoint script for MySQL Server 5.7.44-1.el7 started.
2025-08-27 15:46:48+08:00 [ERROR] [Entrypoint]: mysqld failed while attempting to check configcommand was: mysqld --verbose --help --log-bin-index=/tmp/tmp.lvQBb4Rtwb

进入到容器手动运行mysqld命令也没有有用的信息，会在光标卡一段时间后提示killed。

查找资料

很幸运谷歌到了一个issues：官方docker github里的issuse

里面的回复：

Hello,I know this post is old, nevertheless, for those who come across it in the future like me, this solution allows resolving the issue without intervening in the host system configuration:---
version: '3.6'
services:mysql:image: mysql:5.7ulimits:nofile: # Fix memory leak issue on some systems when LimitCORE=infinity (containerd)soft: 1048576hard: 1048576
This is a configuration for Compose (Docker plugin). However, you can use the equivalent, such as docker run mysql:5.7 --ulimit nofile=1048576:1048576.For your information, I encountered a similar issue with the library/elasticsearch:2.4-alpine image.I use Fedora Workstation Silverblue 39

我原来运行的命令：

docker run -d --name mysql \
-v /app/mysql/binlogs:/ava_app/mysql/binlogs \
-v /app/mysql/conf:/etc/mysql \
-v /app/mysql/logs:/var/log/mysql \
-v /data/upload:/ava_data/upload \
-v /etc/localtime:/etc/localtime:ro \
-v /app/timezone:/etc/timezone:ro \
-e "MYSQL_ROOT_PASSWORD=xxxxxxxxxxxx" \
--restart always --network=host mysql:5.7.44

根据上面方法改进的命令：

# 加入 --ulimit nofile=1048576:1048576 
docker run -d --ulimit nofile=1048576:1048576 --name mysql \
-v /app/mysql/binlogs:/ava_app/mysql/binlogs \
-v /app/mysql/conf:/etc/mysql \
-v /app/mysql/logs:/var/log/mysql \
-v /data/upload:/ava_data/upload \
-v /etc/localtime:/etc/localtime:ro \
-v /app/timezone:/etc/timezone:ro \
-e "MYSQL_ROOT_PASSWORD=xxxxxxxxxxxx" \
--restart always --network=host mysql:5.7.44