[ https://issues.apache.org/jira/browse/MESOS-7652?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Vinod Kone updated MESOS-7652: ------------------------------ Sprint: Mesosphere Sprint 60, Mesosphere Sprint 61 (was: Mesosphere Sprint 60) > Docker image with universal containerizer does not work if WORKDIR is missing > in the rootfs. > -------------------------------------------------------------------------------------------- > > Key: MESOS-7652 > URL: https://issues.apache.org/jira/browse/MESOS-7652 > Project: Mesos > Issue Type: Bug > Components: containerization > Affects Versions: 1.2.1 > Reporter: michael beisiegel > Assignee: Gilbert Song > Priority: Critical > Labels: mesosphere > > hello, > used the following docker image recently > quay.io/spinnaker/front50:master > https://quay.io/repository/spinnaker/front50 > Here the link to the Dockerfile > https://github.com/spinnaker/front50/blob/master/Dockerfile > and here the source > {color:blue}FROM java:8 > MAINTAINER delivery-engineer...@netflix.com > COPY . workdir/ > WORKDIR workdir > RUN GRADLE_USER_HOME=cache ./gradlew buildDeb -x test && \ > dpkg -i ./front50-web/build/distributions/*.deb && \ > cd .. && \ > rm -rf workdir > CMD ["/opt/front50/bin/front50"]{color} > The image works fine with the docker containerizer, but the universal > containerizer shows the following in stderr. > "Failed to chdir into current working directory '/workdir': No such file or > directory" > The problem comes from the fact that the Dockerfile creates a workdir but > then later removes the created dir as part of a RUN. The docker containerizer > has no problem with it if you do > docker run -ti --rm quay.io/spinnaker/front50:master bash > you get into the working dir, but the universal containerizer fails with the > error. > thanks for your help, > Michael -- This message was sent by Atlassian JIRA (v6.4.14#64029)