故障排查 / 容器无法正常运行
1. Docker Stack 环境中容器反复重启
此问题一般是由于配置、防火墙、各种白名单配置不正确引起。
具体表现为:
- 使用浏览器无法打开页面
- 使用
sudo docker ps -a
命令查看容器列表时,发现重启在不断重启
- 在部署服务器本机使用
curl http://localhost:8088
返回curl: (7) Failed to connect to localhost port 8088: Connection refused
错误
- 日志文件中不断输出错误堆栈信息
可能原因及解决方案:
原因 |
解决方案 |
手工修改过配置但配置存在错误 |
检查修改过的配置文件,检查如 YAML 语法、数据库链接信息是否正确 |
修改配置指定了外部服务器,但实际网络不通 |
检查防火墙、阿里云安全组配置、数据库链接白名单等配置 |
2. Docker Stack 环境中容器不存在
此问题一般是因为运行环境不正确引起。
具体表现为:
- 执行
sudo docker stack ls
可以看到dataflux-func
- 执行
sudo docker ps -a
看不到对应容器
- 执行
sudo docker stack ps dataflux-func --no-trunc
,发现容器状态不正常
可能原因及解决方案:
原因 |
解决方案 |
系统中安装的是 snap 版 Docker |
卸载 snap 版 Docker,重新安装官方途径的 Docker,或使用脚本自带的 Docker |
其他 |
可根据sudo docker stack ps dataflux-func --no-trunc 中ERROR 栏目排查 |
典型例子如下,因为磁盘空间不足无法启动:
3. k8s 环境中容器无法启动
此问题一般是因为宿主机 / k8s 集群问题引起。
k8s 中可能存在如下错误:
Text Only |
---|
| Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning Failed 36m kubelet Error: failed to start container "func-server": Error response from daemon: OCI runtime create failed: container_linux.go:367: starting container process caused: process_linux.go:495: container init caused: rootfs_linux.go:60: mounting "/home/cce/kubelet/pods/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/volume-subpaths/user-config/func-server/1" to rootfs at "/home/cce/docker/overlay2/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx/merged/data/user-config-template.yaml" caused: no such file or directory: unknown
Warning Failed 36m kubelet Error: failed to start container "func-server": Error response from daemon: OCI runtime create failed: container_linux.go:367: starting container process caused: process_linux.go:495: container init caused: rootfs_linux.go:60: mounting "/home/cce/kubelet/pods/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/volume-subpaths/user-config/func-server/1" to rootfs at "/home/cce/docker/overlay2/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx/merged/data/user-config-template.yaml" caused: no such file or directory: unknown
Warning Failed 36m kubelet Error: failed to start container "func-server": Error response from daemon: OCI runtime create failed: container_linux.go:367: starting container process caused: process_linux.go:495: container init caused: rootfs_linux.go:60: mounting "/home/cce/kubelet/pods/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/volume-subpaths/user-config/func-server/1" to rootfs at "/home/cce/docker/overlay2/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx/merged/data/user-config-template.yaml" caused: no such file or directory: unknown
Normal Created 35m (x5 over 118d) kubelet Created container func-server
Warning Failed 35m kubelet Error: failed to start container "func-server": Error response from daemon: OCI runtime create failed: container_linux.go:367: starting container process caused: process_linux.go:495: container init caused: rootfs_linux.go:60: mounting "/home/cce/kubelet/pods/xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx/volume-subpaths/user-config/func-server/1" to rootfs at "/home/cce/docker/overlay2/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx/merged/data/user-config-template.yaml" caused: no such file or directory: unknown
Normal Pulled 33m (x6 over 118d) kubelet Container image "harbor.xxx.com/obs/dataflux/dataflux-func:2.7.0" already present on machine
Warning BackOff 2m8s (x157 over 36m) kubelet Back-off restarting failed container
|
某个 Func 服务中可能存在如下错误:
Text Only |
---|
| Traceback (most recent call last):
File "_config.py", line 11, in <module>
CONFIG = yaml_resource.load_config(os.path.join(BASE_PATH, './config.yaml'))
File "/usr/src/app/worker/utils/yaml_resources.py", line 83, in load_config
user_config_content = _f.read()
OSError: [Errno 5] Input/output error
|
这不是 DataFlux Func 的问题,请检查宿主机 / k8s 集群,如涉及 NAS 的使用,也应检查 NAS 是否存在问题。