Details
-
Bug
-
Resolution: Fixed
-
Highest
Description
This just happened to several of our builds - instead of launching in our docker image, the build process was started in the parent kubernetes pod:
+ umask 000 + git clone --branch="develop" --depth 50 https://x-token-auth:$REPOSITORY_OAUTH_ACCESS_TOKEN@bitbucket.org/simpletechs/some-repo.git $BUILD_DIR ; git reset --hard some-hash ; git remote set-url origin https://x-token-auth:{access_token}@bitbucket.org/simpletechs/some-repo.git Cloning into '/opt/atlassian/pipelines/agent/build'... HEAD is now at some-hash some-text + chmod 777 $BUILD_DIR + mkdir tmp + some_command /opt/atlassian/pipelines/agent/tmp/bashScript39....sh: line 16: some_command: command not found
We were able to verify by running a couple of commands that this is, in fact, not our docker image:
# cat /proc/version
Linux version 4.7.3-coreos-r2 (jenkins@worker-1) (gcc version 4.9.3 (Gentoo Hardened 4.9.3 p1.5, pie-0.6.4) ) BCLOUD-1 SMP Thu Feb 2 02:26:10 UTC 2017
# ps auxw
PID USER TIME COMMAND
1 root 0:00 /bin/sh -c exit $(mkfifo /opt/atlassian/pipelines/agent/tmp/build_result && cat /opt/atlassian/pipelines/agent/tmp/build_result)
7 root 0:00 cat /opt/atlassian/pipelines/agent/tmp/build_result
9 root 0:00 /bin/sh /opt/atlassian/pipelines/agent/tmp/wrapperShellScript2853231412848204416.sh
15 root 0:00 /bin/bash /opt/atlassian/pipelines/agent/tmp/bashScript4045879203974059494.sh
16 root 0:00 bash -i
27 root 0:00 ps auxw
This is obviously a rather big deal, as our build require tools that are not present in the parent pod (and why should they be).