VCI allows you to obtain process core dump files, which enables you to analyze and locate container exceptions. This topic mainly describes how to achieve the persistent storage of container core dump files in VCI.

Background information

The core dump is the process in which the operating system saves the current memory state of a program into a core file when the program abnormally terminates due to a runtime exception. If a VCI container abnormally terminates and the application log does not provide enough information to locate the root cause, you need to use the core dump file for further analysis and troubleshooting. This topic describes how to enable containers to generate core dump files and save them to external storage.

Feature introduction

By default, core dumps are disabled for VCIs to prevent excessive disk usage that can lead to service unavailability. You can enable core dumps by using pod annotations when creating a VCI.

Annotation Key	Annotation sample value	Description	Required
vci.vke.volcengine.com/core-pattern	"/data/coredump/core.%h.%e.%p.%t"	Specify the path where core dump files are saved for the VCI. After you specify the path, the core dump capability of the VCI is automatically enabled. Sample value format: /data/coredump/core.%h.%e.%p.%t. The value description is as follows: /data/coredump/: The directory for saving the core dump file. Replace the value with your own directory. core.%h.%e.%p.%t: The name of the core dump file. %h, %e, %p, and %t are placeholders, which you can configure based on your business requirements. The placeholder descriptions are as follows: %h: The name of the VCI pod (hostname). This configuration is recommended. %e: The program file name. This configuration is recommended. %p: The process ID. This configuration is optional. %t: The time of the core dump. This configuration is optional. Warning The configured path cannot start with \|, which indicates that you cannot use a core dump to configure an executable program.	Required in this scenario

The complete path example for the final generated core dump file is as follows:

Prerequisites

A cluster that uses VCI has been created. For details, see Using VCI in VPC-CNI clusters and Using VCI in Flannel clusters.

Procedure

Core dump files are typically used for offline issue analysis. Therefore, when you configure a path for saving core dump files, we recommend that you use attached storage rather than built-in storage (local container paths). This prevents the loss of core dump files when the container exits.

The following example demonstrates how to enable the core dump capability for a VCI Deployment by using FileNAS as attached storage. This is achieved by mounting FileNAS to the VCI Deployment via the PV method. For detailed instructions and YAML field descriptions, see Using FileNAS as static PersistentVolumes.

Connect to the VPC-CNI cluster.

For details, see Connecting to clusters.

Create a PV of the FileNAS type.

Create a YAML file for the PV. The following vci-nas-pv.yaml file serves as an example:

apiVersion: v1
kind: PersistentVolume
metadata:
name: vci-nas-pv # The name of the PV.
spec:
accessModes:
- ReadWriteMany
capacity:
storage: 100Gi # FileNAS storage capacity.
csi:
driver: nas.csi.volcengine.com
volumeAttributes:
fsId: enas-cnbj1b9f******** # The ID of the FileNAS file system. This parameter needs to be specified only if the file system type is Extreme speed.
fsType: Extreme # File system type. Valid values include Extreme, Cache, and Capacity.
path: /enas-cnbj1b9f********/temp # The subdirectory of the FileNAS file system for mounting.
server: cnbj1b9f********.r*********.nas.ivolces.com # The mount point of the FileNAS file system.
volumeAs: subpath # The subdirectory type. Valid values include subpath and filesystem.
volumeHandle: vci-nas-pv # The unique identifier of the PV. Keep it consistent with the PV name.
mountOptions:
- nolock,proto=tcp,noresvport
- vers=3
persistentVolumeReclaimPolicy: Retain
volumeMode: Filesystem

Run the following command to create the PV.

kubectl create -f vci-nas-pv.yaml

Create a PersistentVolumeClaim (PVC).

Create a YAML file for the PVC. The following vci-nas-pvc.yaml file serves as an example:

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: vci-nas-pvc # The name of the PVC.
namespace: default
spec:
accessModes:
- ReadWriteMany
resources:
requests:
storage: 100Gi
volumeMode: Filesystem
volumeName: vci-nas-pv # The name of the PV bound to this PVC.

Run the following command to create the PVC.

kubectl create -f vci-nas-pvc.yaml

Create a VCI, configure a path for storing core dump files, and enable core dumps for the VCI.

The following content provides an example of how to create a VCI Deployment. For instructions on creating a VCI, see Creating instances.

Create a YAML file for the VCI Deployment. The sample YAML file vci-nas-coredump-deployment.yaml is as follows:

apiVersion: apps/v1
kind: Deployment
metadata:
name: vci-nas-coredump-deployment # The name of the Deployment.
namespace: default # The namespace of the Deployment.
labels:
app: vci-nas-coredump-deployment
spec:
replicas: 2
selector:
matchLabels:
app: vci-nas-coredump-deployment
template:
metadata:
annotations:
vke.volcengine.com/burst-to-vci: enforce # Force the use of VCI.
vci.vke.volcengine.com/preferred-instance-family: vci.u1 # Specify the instance family of VCI.
vci.vke.volcengine.com/core-pattern: "/data/coredump/core.%h.%e.%p.%t" # Specify the path where the core dump file is saved.
labels:
app: vci-nas-coredump-deployment
spec:
containers:
- image: cr-cn-beijing.volces.com/tw/nginx:1.0 # The address and version of the container image. Replace the value with your actual information.
imagePullPolicy: IfNotPresent
name: nginx
volumeMounts:
- mountPath: /data/coredump/ # The location where the FileNAS storage is mounted in the container. This must be consistent with the path specified in this topic where the core dump file is saved.
name: vci-nas-volume # Specify the volume name.
volumes:
- name: vci-nas-volume # The mount volume.
persistentVolumeClaim:
claimName: vci-nas-pvc # The name of the PVC created in the previous step.

Run the following command to create the Deployment.

kubectl create -f vci-nas-coredump-deployment.yaml

Validating results

You can trigger a core dump in a VCI for which a directory for saving core dump files is configured.

Run the following command to obtain the name of the pod in the VCI Deployment.

kubectl get pods -l app=vci-nas-coredump-deployment

Run the following command to log in to the VCI pod.

kubectl exec -it ${the obtained VCI Pod name}

Run the sleep 100 command in the container and then press Ctrl+\ to trigger a core dump. The generated core dump file will be automatically saved to FileNAS.

The expected response is as follows:

Volcengine Kubernetes Engine