[builds][Conformance] build without output image

[0xmichalis]

• Failure [39.765 seconds]
[builds][Conformance] build without output image
/data/src/github.com/openshift/origin/test/extended/builds/no_outputname.go:49
  building from templates
  /data/src/github.com/openshift/origin/test/extended/builds/no_outputname.go:48
    should create an image from "/data/src/github.com/openshift/origin/test/extended/testdata/test-s2i-no-outputname.json" S2i template without an output image reference defined [It]
    /data/src/github.com/openshift/origin/test/extended/builds/no_outputname.go:47

    Expected error:
        <*exec.ExitError | 0xc8206a47a0>: {
            ProcessState: {
                pid: 25612,
                status: 256,
                rusage: {
                    Utime: {Sec: 0, Usec: 113776},
                    Stime: {Sec: 0, Usec: 19652},
                    Maxrss: 25448,
                    Ixrss: 0,
                    Idrss: 0,
                    Isrss: 0,
                    Minflt: 6455,
                    Majflt: 0,
                    Nswap: 0,
                    Inblock: 0,
                    Oublock: 0,
                    Msgsnd: 0,
                    Msgrcv: 0,
                    Nsignals: 0,
                    Nvcsw: 2150,
                    Nivcsw: 22,
                },
            },
            Stderr: nil,
        }
        exit status 1
    not to have occurred

    /data/src/github.com/openshift/origin/test/extended/builds/no_outputname.go:45

@stevekuznetsov known flake?

[0xmichalis]

https://ci.openshift.redhat.com/jenkins/job/test_pull_requests_origin_conformance/2426/consoleFull

[smarterclayton]

@bparees didn't we fix these flakes?

[bparees]

@smarterclayton this looks different, looks like docker daemon wackiness. from the build log:

I0622 13:45:23.317312       1 glog.go:50] Removing container 
"20aa7dc686459824a457d6cb01e1e84cce2365ab76912178ff0927c7efa45043" ...
warning: Failed to remove container "20aa7dc686459824a457d6cb01e1e84cce2365ab76912178ff0927c7efa45043": API error (500): Driver devicemapper failed to remove root filesystem 20aa7dc686459824a457d6cb01e1e84cce2365ab76912178ff0927c7efa45043: mount still active
I0622 13:45:23.390875       1 cleanup.go:31] Removing temporary directory /tmp/s2i-build031277940
I0622 13:45:23.390901       1 fs.go:156] Removing directory '/tmp/s2i-build031277940'
F0622 13:45:23.399964       1 helpers.go:107] error: build error: building extended-test-build-no-outputname-yeaf0-xyuep/test-sti-1:72737d39 failed when committing the image due to error: No such container: 20aa7dc686459824a457d6cb01e1e84cce2365ab76912178ff0927c7efa45043
error: The build extended-test-build-no-outputname-yeaf0-xyuep/test-sti-1 status is "Failed"

[stevekuznetsov]

@Kargakis new flake? Have you added a Jenkins Failure Cause Management entry?

[stevekuznetsov]

Devicemapper issues... @runcom @rhvgoyal maybe you will be able to shed some light on it?

[bparees]

per my comment above, this is a docker daemon problem

[bparees]

recurred:
https://ci.openshift.redhat.com/jenkins/job/test_pull_requests_origin_conformance/2849/consoleFull

[bparees]

https://ci.openshift.redhat.com/jenkins/job/test_pull_requests_origin_conformance/3764/consoleFull

[stevekuznetsov]

@runcom @rhvgoyal any thoughts on this one?

[mfojtik]

bumping prio of this as this is no. 1 flake :-)

[bparees]

@mfojtik it's number one only because it's been around so long, if you look at the actual frequency with which it's hit, it's not very frequent (hasn't been seen in 15 days for instance).

Need to divide number of occurences by the age of the flake, to get at least a simplistic metric.

[mfojtik]

@bparees good idea :-) i will add that tonight to my flake reporter.

[jwhonce]

@mfojtik @bparees Given your discussion should this really be restored to a P2?

[bparees]

@jwhonce given that it's been 25+ days since anyone referenced it, i'd even be willing to consider closing it and reopen if we recreate, unless you guys have reason to think it's still a problem.

[jwhonce]

Please re-open issue if it resurfaces.

[stevekuznetsov]

Rats. We've not been referencing this but we hit it in base_ami job quite often: https://ci.openshift.redhat.com/jenkins/job/base_ami/2571/consoleFull#1518085191577f0ce7e4b0b14b5836ce6d

That being said this feels like a weird conglomerate of lots of issues, but @runcom should be able to better understand it.

[runcom]

Pretty sure there's an active bugzilla about this "mount still active" on Docker build (https://bugzilla.redhat.com/show_bug.cgi?id=1403027)