Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-51133

Branch names in Multibranch jobs causing build failures

    Details

    • Similar Issues:

      Description

      Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

      Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

      We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

      [ERROR] COMPILATION ERROR : 
      18:16:23 [INFO] -------------------------------------------------------------
      18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
      18:16:23   symbol:   class Consumer
      18:16:23   location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
      18:16:23 [INFO] 1 error
      

      The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same physical host, at separate times.

      The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

      Deleting the branch, removing it from the multibranch job, then re-creating the branch does not fix the issue

      The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

      The only correlation I can see is with the branch name. We occasionally see this with PRs too - sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code change fixes the problem

        Attachments

          Activity

          robin_smith Robin Smith created issue -
          robin_smith Robin Smith made changes -
          Field Original Value New Value
          Description Occasionally, forking a new branch in a git repo can cause immediate, unrelated build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too. Sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code changes fixes the problem
          Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too. Sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code changes fixes the problem
          robin_smith Robin Smith made changes -
          Summary Branch names causing build failures Branch names in Multibranch jobs causing build failures
          robin_smith Robin Smith made changes -
          Description Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too. Sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code changes fixes the problem
          Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same physical host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too. Sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code changes fixes the problem
          robin_smith Robin Smith made changes -
          Description Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same physical host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too. Sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code changes fixes the problem
          Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same physical host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too - sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code changes fixes the problem
          robin_smith Robin Smith made changes -
          Description Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same physical host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too - sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code changes fixes the problem
          Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same physical host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too - sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code change fixes the problem
          robin_smith Robin Smith made changes -
          Description Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same physical host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too - sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code change fixes the problem
          Occasionally, forking a new branch in a git repo can cause immediate, unrelated and irreversible build failures.

          Our setup is a multibranch pipeline building each branch of the project for every commit. The repo is stored in Bitbucket.

          We recently created a 'release/6.0.0' branch from our master branch, and the build on that branch immediately failed with the following error:

          {code}
          [ERROR] COMPILATION ERROR :
          18:16:23 [INFO] -------------------------------------------------------------
          18:16:23 [ERROR] /root/workspace/enAM-Pipeline_release_6.0.0-YRZAEYTMSVCBGF6L6YJT7VNUKLPLU3KCUEIOE5P5NURDINYXJM4Q/openam-core/src/main/java/org/forgerock/openam/entitlement/SetupInternalNotificationSubscriptions.java:[55,73] cannot find symbol
          18:16:23 symbol: class Consumer
          18:16:23 location: class org.forgerock.openam.entitlement.SetupInternalNotificationSubscriptions
          18:16:23 [INFO] 1 error
          {code}

          The same commit built fine on master, the only change in git is the creation of this new branch, which gets a new job/branch in the multibranch pipeline to reflect that. Our builds run on fresh Docker containers (Yet Another Docker Plugin); the builds in question ran on the same physical host, at separate times.

          The problem is not ephemeral for the new branch - all future builds of that branch will be broken, with the same issue.

          Deleting the branch, removing it from the multibranch job, then re-creating the branch does not fix the issue

          The only approach that seems to work is to create another branch directly from the 'broken' one, with another name. For example, we created 'release/6.0.0b' (directly from 'release/6.0.0') and the build for that was fine.

          The only correlation I can see is with the branch name. We occasionally see this with PRs too - sometimes the builds for a PR seem irreversibly broken. Simply opening a new PR with the same code change fixes the problem
          abayer Andrew Bayer made changes -
          Component/s bitbucket-branch-source-plugin [ 21428 ]
          Component/s workflow-multibranch-plugin [ 21465 ]
          Component/s pipeline [ 21692 ]
          miraha jang hyemi (Inactive) made changes -
          Status Open [ 1 ] In Progress [ 3 ]
          miraha jang hyemi (Inactive) made changes -
          Status In Progress [ 3 ] In Review [ 10005 ]
          vivek Vivek Pandey made changes -
          Labels multibranch multibranch triaged-2018-11
          vivek Vivek Pandey made changes -
          Status In Review [ 10005 ] In Progress [ 3 ]
          vivek Vivek Pandey made changes -
          Status In Progress [ 3 ] Open [ 1 ]
          Hide
          vivek Vivek Pandey added a comment -

          Switching back to Open from In Review since no link to PR.

          Show
          vivek Vivek Pandey added a comment - Switching back to Open from In Review since no link to PR.
          Hide
          robin_smith Robin Smith added a comment -

          Not seen in almost a year - closing

          Show
          robin_smith Robin Smith added a comment - Not seen in almost a year - closing
          robin_smith Robin Smith made changes -
          Status Open [ 1 ] Closed [ 6 ]
          Resolution Fixed [ 1 ]

            People

            • Assignee:
              Unassigned
              Reporter:
              robin_smith Robin Smith
            • Votes:
              1 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: