Goal: Allow Quay administrators to supplement persistent storage to managed Clair deployments to provide a bigger scratch space for image layer extraction.

Background: We've had customers in the past who were exposing Clair to unexpectedly large images (03431241). These large images and numerous past issues around cleaning up the temp directories used for extracting image layers (PROJQUAY-5119) can cause the temporary directories to require more disk space than the host can offer. While we eventually need to fix the bugs that cause extracted image layers to remain in temporary directories after they are no longer needed, an escape valve by providing larger directories is beneficial to prevent an outage of Clair.

Requirements:

• the Quay operator allows users to configure a persistent volume to be mounted to the Clair instances, supplementing the /tmp directory
• Users should be able to specify desired capacity and storage class of the PersistentVolume at the level of the managed clair component
• Adhere to Kubernetes Best Practices when working with PersistentVolumeClaims: https://kubernetes.io/docs/concepts/storage/persistent-volumes/#writing-portable-configuration