Uploaded image for project: 'OpenShift Request For Enhancement'
  1. OpenShift Request For Enhancement
  2. RFE-5232

Improve cri-o Wipe Storage Corruption Protection

XMLWordPrintable

    • Icon: Feature Request Feature Request
    • Resolution: Done
    • Icon: Undefined Undefined
    • None
    • openshift-4.12.z, openshift-4.14.z
    • Node
    • False
    • None
    • False
    • Not Selected
    • OCPSTRAT-1291 - Phase -1 : Improve CRIO wipe for reboot

      1. Proposed title of this feature request

         Improve CRI-O Wipe Storage Corruption Protection

      2. What is the nature and description of the request?

      On an unplanned or unexpected reboot of a node `crio-wipe` runs to clean up the storage to avoid having corrupted images. However cleaning all images seems quite invasive and is extending startup time. Further on it requires that the registry (internal/or external) is available when the nodes comes up again which may esp in telco edge locations not necessarily be the case.

      Hence it would be good if cri-o would have a possibility to detect only corrupted layers and wipe them instead of all images which would help keeping the time for wipe and reboot at a minimum.

      3. Why does the customer need this? (List the business requirements here)

      Telco edge locations are often having just minimal set up with regards to data center technology and at times power availability may not be guaranteed. In such cases unexpected reboots (power spike) or outage are much more common than in other data center location. Hence it is needed to have this process improved to achieve better treatment of corrupted images and improve reboot times in case layers need to be pruned and repulled

       

              gausingh@redhat.com Gaurav Singh
              rhn-support-dmoessner Daniel Moessner
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

                Created:
                Updated:
                Resolved: