Docker MySQL 容器故障修复记录

环境

容器：1Panel-mysql-vVOO
镜像：mysql:8.4.4
Compose：/opt/1panel/apps/mysql/mysql/docker-compose.yml
配置：/opt/1panel/apps/mysql/mysql/conf/my.cnf
数据目录：/opt/1panel/apps/mysql/mysql/data
MySQL 地址：127.0.0.1:30006
修复日期：2026-07-14

不要在此文档或 shell 历史中记录 sudo、MySQL root 等明文密码。以下使用环境变量占位。

bash
export SUDO_PASSWORD='<sudo-password>'
export MYSQL_ROOT_PASSWORD='<mysql-root-password>'
export MYSQL_CONTAINER='1Panel-mysql-vVOO'
export MYSQL_APP='/opt/1panel/apps/mysql/mysql'

故障表现与根因

检查容器：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S docker ps -a
printf '%s\n' "$SUDO_PASSWORD" | sudo -S docker logs --tail 300 "$MYSQL_CONTAINER"

容器以退出码 1 停止。日志包含 InnoDB 断言失败和以下 socket 锁错误：

text
[ERROR] [MY-013183] [InnoDB] Assertion failure
[ERROR] [MY-010276] [Server] Unix socket lock file is empty /var/run/mysqld/mysqld.sock.lock
[ERROR] [MY-010268] [Server] Unable to setup unix socket lock file.

检查容器配置和挂载：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S docker inspect "$MYSQL_CONTAINER" --format '{{json .Config}}\n{{json .HostConfig}}\n{{json .Mounts}}\n{{json .State}}'

原 MySQL 配置包含：

ini
innodb_force_recovery=6

该参数表明 InnoDB 已损坏，6 是最高应急恢复等级，只适用于只读导出，不能用于正常生产运行。旧容器可写层另有残留空 socket 锁文件，导致启动中止。

检查存储与配置

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S df -h / /opt /var/lib/docker
printf '%s\n' "$SUDO_PASSWORD" | sudo -S sh -c 'df -i / /opt /var/lib/docker'

修复时数据目录为约 4.7 GiB，磁盘余量约 283 GiB。

第一次容器重建

为清除旧容器运行层内残留的 socket 锁文件，先强制重建容器。此命令不删除 bind mount 中的数据目录：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S docker compose -f "$MYSQL_APP/docker-compose.yml" up -d --force-recreate

确认 MySQL 仍在应急恢复模式：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S docker exec "$MYSQL_CONTAINER" mysql --protocol=socket -uroot -p"$MYSQL_ROOT_PASSWORD" -e 'SELECT VERSION() AS version, @@innodb_force_recovery AS innodb_force_recovery;'

结果为 innodb_force_recovery = 6。

创建备份

先创建包含权限和属主的物理备份，并备份原配置：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S sh -c 'backup=/opt/1panel/apps/mysql/mysql/data-backup-20260714-0309; test -d /opt/1panel/apps/mysql/mysql && test ! -e "$backup" && cp -a /opt/1panel/apps/mysql/mysql/data "$backup" && cp -a /opt/1panel/apps/mysql/mysql/conf/my.cnf /opt/1panel/apps/mysql/mysql/conf/my.cnf.backup-20260714-0309 && du -sh "$backup"'

在强制恢复模式下导出全部数据库、例程、事件和触发器：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S sh -c 'test -d /opt/1panel/apps/mysql/mysql && test ! -e /opt/1panel/apps/mysql/mysql/recovery-export-20260714-0616.sql && docker exec 1Panel-mysql-vVOO mysqldump --protocol=socket -uroot -p"$MYSQL_ROOT_PASSWORD" --all-databases --routines --events --triggers --hex-blob --single-transaction > /opt/1panel/apps/mysql/mysql/recovery-export-20260714-0616.sql && ls -lh /opt/1panel/apps/mysql/mysql/recovery-export-20260714-0616.sql'

逻辑导出文件为 recovery-export-20260714-0616.sql，大小约 61 MiB。

重建干净的数据库实例

从 my.cnf 中移除：

ini
innodb_force_recovery=6

其余配置保持：

ini
[mysqld]
host-cache-size=0
skip-name-resolve
datadir=/var/lib/mysql
socket=/var/run/mysqld/mysqld.sock
secure-file-priv=/var/lib/mysql-files
user=mysql
pid-file=/var/run/mysqld/mysqld.pid

[client]
socket=/var/run/mysqld/mysqld.sock

!includedir /etc/mysql/conf.d/

停止旧实例，保留旧数据目录，创建新的空数据目录，写入正常配置并重新启动：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S sh -c 'set -e; app=/opt/1panel/apps/mysql/mysql; test -d "$app"; test -f "$app/recovery-export-20260714-0616.sql"; docker compose -f "$app/docker-compose.yml" down; mv "$app/data" "$app/data-before-rebuild-20260714-0619"; mkdir "$app/data"; chown 999:999 "$app/data"; install -o root -g root -m 0644 /tmp/mysql-my.cnf-normal "$app/conf/my.cnf"; docker compose -f "$app/docker-compose.yml" up -d'

等待 MySQL 初始化完成：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S sh -c 'for i in $(seq 1 30); do if docker exec 1Panel-mysql-vVOO mysqladmin --protocol=socket -uroot -p"$MYSQL_ROOT_PASSWORD" ping >/dev/null 2>&1; then exit 0; fi; sleep 10; done; docker logs --tail 120 1Panel-mysql-vVOO; exit 1'

正常结果：

text
mysqld is alive

导入备份

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S sh -c 'set -e; docker exec -i 1Panel-mysql-vVOO mysql --protocol=socket -uroot -p"$MYSQL_ROOT_PASSWORD" < /opt/1panel/apps/mysql/mysql/recovery-export-20260714-0616.sql; docker exec 1Panel-mysql-vVOO mysql --protocol=socket -uroot -p"$MYSQL_ROOT_PASSWORD" -e "SELECT VERSION() AS version, @@innodb_force_recovery AS innodb_force_recovery, @@read_only AS read_only; SHOW DATABASES;"'

导入后应确认：

text
innodb_force_recovery = 0
read_only = 0

功能与重启验证

执行临时读写测试并检查 Nextcloud 表数量：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S docker exec "$MYSQL_CONTAINER" mysql --protocol=socket -uroot -p"$MYSQL_ROOT_PASSWORD" -e "CREATE DATABASE mysql_recovery_verify; CREATE TABLE mysql_recovery_verify.t (id INT PRIMARY KEY); INSERT INTO mysql_recovery_verify.t VALUES (1); SELECT COUNT(*) AS persisted_write_count FROM mysql_recovery_verify.t; DROP DATABASE mysql_recovery_verify; SELECT COUNT(*) AS nextcloud_table_count FROM information_schema.tables WHERE table_schema = 'nextcloud';"

修复时结果：临时写入成功，nextcloud 有 167 张表。

重启后再次验证：

bash
printf '%s\n' "$SUDO_PASSWORD" | sudo -S docker restart "$MYSQL_CONTAINER"

最终验证结果：容器为 `running`，退出码为 `0`，`innodb_force_recovery=0` 且 `read_only=0`。

## 保留的恢复文件

在确认 Nextcloud 长期正常前，不要删除：

```text
/opt/1panel/apps/mysql/mysql/data-backup-20260714-0309
/opt/1panel/apps/mysql/mysql/data-before-rebuild-20260714-0619
/opt/1panel/apps/mysql/mysql/recovery-export-20260714-0616.sql
/opt/1panel/apps/mysql/mysql/conf/my.cnf.backup-20260714-0309