[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: "manifest unknown" test flakes

With the origin image last week, it was persistent until we specified building from the v1.2.0 tag instead of "latest". With this image it has happened twice in a row. I'll try another... https://ci.openshift.redhat.com/jenkins/job/test-origin-aggregated-logging/382/

May be able to reproduce manually, we will see. What would it look like if docker is flaking?

On Tue, Jun 7, 2016 at 9:48 PM, Clayton Coleman <ccoleman redhat com> wrote:
We are building and pushing our images with Docker 1.10 for Origin (but possibly not in all AMIs, depending on the test job).  The last set of pushed images (alpha.1) was pushed from Docker 1.9.

This might just be docker flaking, but we'd probably need more info from the Docker logs.

On Tue, Jun 7, 2016 at 9:41 PM, Luke Meyer <lmeyer redhat com> wrote:
I've seen this twice in logging tests, in the ruby STI builder today and in the origin image last week, so I'm wondering if it's a trend.

Pulling Docker image centos/ruby-22-centos7 sha256:e5ebb014d02b5a7faa7b83f6c58f4cc4eb9892edbf61172e59f2ae1999982dc2 ...
I0608 01:08:34.707737       1 glog.go:50] An error was received from the PullImage call: manifest unknown: manifest unknown

The closest thing I could find under issues is https://github.com/openshift/origin/issues/9122 where it was reportedly due to the image being pushed by docker 1.10.

Is someone pushing our images to dockerhub with docker 1.10?

Builds run on CentOS7, don't think that AMI has docker 1.10 yet.

dev mailing list
dev lists openshift redhat com

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]