Description of problem: for some applications, osc build-logs reports "error: error streaming logs: unexpected EOF", also the build status is Complete Version-Release number of selected component (if applicable): [beta4/2015-06-04.2] How reproducible: 100% Steps to Reproduce: 1. osc new-app https://github.com/anpingli/example-perl 2. osc start-build example-perl 3. osc build-logs example-perl-1 4. osc get build Actual results: For step 3. osc build-logs example-perl-1 <---snip---> Removing intermediate container 4b21daa529ad Step 5 : ENV OPENSHIFT_BUILD_SOURCE "https://github.com/anpingli/example-perl" OPENSHIFT_BUILD_NAME "example-perl-1" OPENSHIFT_BUILD_NAMESPACE "p2" ---> Running in 5e9a35b5a089 ---> 729825eddfd7 Removing intermediate container 5e9a35b5a089 Successfully built 729825eddfd7 I0607 09:38:24.569588 1 cfg.go:50] Problem accessing /root/.dockercfg: stat /root/.dockercfg: no such file or directory I0607 09:38:24.569643 1 docker.go:91] Pushing 172.30.134.207:5000/p2/example-perl:latest image ... error: error streaming logs: unexpected EOF [joe@master ~]$ echo $? 1 For step 4: [joe@master ~]$ osc get build NAME TYPE STATUS POD example-perl-1 Docker Complete example-perl-1-build Expected results: osc build-logs shouldn't end with unexpected EOF Additional info:
I suspect it happened after 5 minutes. This is likely the same as: https://github.com/GoogleCloudPlatform/kubernetes/issues/9013 https://github.com/openshift/origin/issues/2348 I think the right solution would be for the client to reconnect and pick up where it left off. This won't happen for 3.0 though.
I'm fairly certain this has been worked around by increasing the timeout. If you still hit this problem please reopen this bug.