OpenShift 3.6 Issues with Hard-Pruning

Solution In Progress - Updated -

Issue

During hard pruning following the documentation [1] 2 issues preventing the operation to succeed

Issue 1:
1107 17:53:15.502689 235 round_trippers.go:414] Content-Type: application/json
I1107 17:53:15.502699 235 round_trippers.go:414] Content-Length: 248
I1107 17:53:15.502708 235 round_trippers.go:414] Date: Wed, 07 Nov 2018 17:53:15 GMT
I1107 17:53:15.502718 235 round_trippers.go:414] Cache-Control: no-store
I1107 17:53:15.502746 235 request.go:991] Response Body: {"kind":"Status","apiVersion":"v1","metadata":{},"status":"Failure","message":"imagestreams.image.openshift.io \"devimage\" not found","reason":"NotFound","details":{"name":"devimage","group":"image.openshift.io","kind":"imagestreams"},"code":404}
I1107 17:53:15.509421 235 request.go:782] Error in request: invalid resource name "dev-project-exporter/dev-project/dev_projectexporter_java": [may not contain '/']
ERRO[0338] failed to get the image stream certificates/ev-project-exporter/dev-project/dev_projectexporter_java: invalid resource name "ev-project-exporter/dev-project/dev_projectexporter_java": [may not contain '/']
Would delete 0 blobs

In 3.9 we added a restriction so you can't upload blobs into the repository with an invalid name: bug 1528613.
The registry stores repositories in /docker/registry/v2/repositories/. It is safe to remove repositories with invalid names from this directory.

Issue 2:
1120 16:43:42.473167 721 request.go:991] Response Body: {"kind":"ImageStream","apiVersion":"v1","metadata":{"name":"openshift-hello","namespace":"monitoring-clustercheck","selfLink":"/oapi/v1/namespaces/monitoring-clustercheck/imagestreams/openshift-hello","uid":"4f3845aa-ece3-11e8-83ad-40f2e91e200b","resourceVersion":"582035967","generation":1,"creationTimestamp":"2018-11-20T16:42:46Z","labels":{"app":"openshift-hello"},"annotations":{"openshift.io/generated-by":"OpenShiftNewApp"}},"spec":{"lookupPolicy":{"local":false}},"status":{"dockerImageRepository":"173.20.149.84:5000/monitoring-clustercheck/openshift-hello","tags":[{"tag":"latest","items":[{"created":"2018-11-20T16:43:00Z","dockerImageReference":"173.20.149.84:5000/monitoring-clustercheck/openshift-hello@sha256:0bd4a9f017e91e8f7621a85f7a355025a7125ccb5cc5f79e36b7ee9b04107857","image":"sha256:0bd4a9f017e91e8f7621a85f7a355025a7125ccb5cc5f79e36b7ee9b04107857","generation":1}]}]}}
ERRO[0819] failed to prune manifest links in the repository monitoring-clustercheck/openshift-hello: invalid checksum digest format

[1]
https://access.redhat.com/documentation/en-us/openshift_container_platform/3.6/html/cluster_administration/admin-guide-pruning-resources#hard-pruning-registry

Environment

OpenShift 3.6 latest patchlevel
Previously normal pruning was not working because ETCD timeout. To fix fix many images needed to be deleted via oc delete image leaving the registry in an inconsistent state making hard-pruning necessary

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.

Current Customers and Partners

Log in for full access

Log In

New to Red Hat?

Learn more about Red Hat subscriptions

Using a Red Hat product through a public cloud?

How to access this content