Skip to content

container event can't be handled by containerd when restart containerd  #10390

@ningmingxiao

Description

@ningmingxiao

Description

$ cat eventlost.txt |grep 56e3c55053e89ca9e
Jun 24 11:04:35 icf-test1-minion-0-0 containerd[235439]: time="2024-06-24T11:04:35.913762399+08:00" level=info msg="Container 56e3c55053e89ca9eea56382d0e11929db019098a27ee8e685cca55f7ebf6807: CDI devices from CRI Config.CDIDevices: []"
Jun 24 11:04:35 icf-test1-minion-0-0 containerd[235439]: time="2024-06-24T11:04:35.965088489+08:00" level=info msg="CreateContainer within sandbox "6d243882bc36f1b41f2b07208e087e24320bc0bf354ede8f6f858e3a9a9dacb5" for &ContainerMetadata{Name:init-process-ctrl-processd,Attempt:0,} returns container id "56e3c55053e89ca9eea56382d0e11929db019098a27ee8e685cca55f7ebf6807""
Jun 24 11:04:35 icf-test1-minion-0-0 containerd[235439]: time="2024-06-24T11:04:35.965645896+08:00" level=info msg="StartContainer for "56e3c55053e89ca9eea56382d0e11929db019098a27ee8e685cca55f7ebf6807""
Jun 24 11:04:36 icf-test1-minion-0-0 containerd[235439]: time="2024-06-24T11:04:36.197516307+08:00" level=info msg="StartContainer for "56e3c55053e89ca9eea56382d0e11929db019098a27ee8e685cca55f7ebf6807" returns successfully"
Jun 24 11:04:41 icf-test1-minion-0-0 containerd[566974]: time="2024-06-24T11:04:41.391999520+08:00" level=info msg="received exit event container_id:"56e3c55053e89ca9eea56382d0e11929db019098a27ee8e685cca55f7ebf6807" id:"56e3c55053e89ca9eea56382d0e11929db019098a27ee8e685cca55f7ebf6807" pid:564774 exited_at:{seconds:1719198281 nanos:391318113}"

container is exited, crictl show error status
nmx888

Steps to reproduce the issue

37b0b9d6b945d7750dea10924137957a525e6108b9a279077566d708dae78f23 is an already running cri container.
1.add some delay to make it easier to reproduce in
func (c *criService) recover(ctx context.Context) error

			cntr, err := c.loadContainer(ctx2, container)
			if container.ID() == "37b0b9d6b945d7750dea10924137957a525e6108b9a279077566d708dae78f23" {
				fmt.Println("*******kill *************")
				time.Sleep(time.Second * 30)
			}

2.compile and run containerd,see *****kill *** log ,then kill 37b0b9d6b945d7750dea10924137957a525e6108b9a279077566d708dae78f23 main process in 30s
3.wait containerd started, crictl ps show container is still runnning

Describe the results you received and expected

crictl show container is exited

What version of containerd are you using?

1.7

Any other relevant information

No response

Show configuration if it is related to CRI plugin.

No response

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    Projects

    Status

    Done

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions