Description
When the CSI driver invokes the func (resizefs *ResizeFs) Resize(devicePath string, deviceMountPath string) (bool, error) function, it is possible for the execution of the resize2fs command to take longer than the timeout specified for Kubernetes CSI interface calls. This scenario results in the CSI interface returning a failure, even though the resize2fs command is still in progress. Consequently, if Kubernetes attempts to call the CSI interface again, it may inadvertently issue another resize2fs command.
To mitigate this issue, could we implement a "Resize method with timeout handling"? This would allow the CSI driver to proactively return an error during the interface call, rather than allowing Kubernetes to time out and assume the operation has failed.