cdk-addons build doesn't log docker pull failures
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| Charmed Kubernetes Testing |
Triaged
|
Low
|
Kevin W Monroe | ||
Bug Description
When docker pull fails for some reason, the error is not captured and reported. It later shows up as a failure when trying to push the image into rocks.
An example from a failed run:
01:06:05 + docker pull quay.io/
01:06:05 + grep -qi no matching manifest for
01:06:27 + RAW_IMAGE=
01:06:27 + grep -qi docker.io/
01:06:27 + echo quay.io/
01:06:27 + grep -qi k8s.gcr.io/
01:06:27 + echo quay.io/
01:06:27 + grep -qi quay.io/
01:06:27 + echo quay.io/
01:06:27 + sed -e s|quay.io/||g
01:06:27 + echo quay.io/
01:06:27 + RAW_IMAGE=
01:06:27 + break
01:06:27 + grep -qi coredns
01:06:27 + echo cephcsi/
01:06:27 + grep -qi k8s-dns-kube-dns
01:06:27 + echo cephcsi/
01:06:27 + echo cephcsi/
01:06:27 + grep -qi k8s-dns-
01:06:27 + echo cephcsi/
01:06:27 + grep -qi k8s-dns-sidecar
01:06:27 + docker tag quay.io/
01:06:27 Error response from daemon: No such image: quay.io/
This makes this type of error difficult to track down. Errors should somehow be captured from https:/
| Changed in charmed-kubernetes-testing: | |
| assignee: | nobody → Kevin W Monroe (kwmonroe) |
| status: | New → Confirmed |
| importance: | Undecided → High |
| Changed in charmed-kubernetes-testing: | |
| importance: | High → Medium |
| Changed in charmed-kubernetes-testing: | |
| importance: | Medium → Low |
| status: | Confirmed → Triaged |
