Storing coredumps generated by different processes having different UIDs in Azure File Persistent Volume

Question

As part of my K8S cluster, i'm using AzureFile PV to store coredumps generated by different processes running as part of my application (I'm using Azure File as, i need to access the volume from different containers running in multiple worker nodes/VMs).

The issue that i'm facing is that, the coredump generated by the application is non-zero, if the UID of the process is root. The coredumps genretaed by application is of zero size, if the UID of the process is non-root.

Following is my storage class

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: {{ .Release.Name }}-azurefile-core-sc
namespace: {{ .Values.global.namespace }}
provisioner: kubernetes.io/azure-file
mountOptions:

• dir_mode=0777
• file_mode=0600
• uid=root
• gid=root
• mfsymlinks
• cache=strict
  parameters:
  skuName: Standard_LRS

Answer 1

@Shanmugam, Magesh
I believe this issue could be caused because you have file_mode=0600. This will only allow the root user to read or write files. Can you try setting this to file_mode=0777 to see if that resolves your issue?

Hope this helps. If not please let me know the results and we will continue to investigate from our side.

-------------------------------

Please don’t forget to "Accept the answer" and “up-vote” wherever the information provided helps you, this can be beneficial to other community members.