• Icon: Bug Bug
    • Resolution: Fixed
    • Icon: Major Major
    • core
    • None
    • Platform: All, OS: All

      For no reason that I can see,
      http://deadlock.netbeans.org/hudson/job/nbms-and-javadoc/3716/ ran in a fresh
      workspace:

      A SCM change trigger started this job
      Building on master
      $ hg clone http://hg.netbeans.org/main-silver/
      /hudson/workdir/jobs/nbms-and-javadoc/workspace@2
      ....

      The #3715 and #3717 used the regular workspace (and thus could use hg update
      rather than the far slower clone).
      http://deadlock.netbeans.org/hudson/computer/(master)/builds confirms that all
      three ran on the master node.

      I looked for code in main/core/src/main/java that would create such a directory
      but could not find it.

          [JENKINS-4202] Build run on .../workspace@2

          I suspect this is caused by the SCM polling thread using the workspace just when
          the next build was about to start. I'll write a test case to examine this behavior.

          I assume the previous behavior of blocking the build until the polling is over
          is desirable, right? At least unless the "do concurrent build" option is checked.

          Kohsuke Kawaguchi added a comment - I suspect this is caused by the SCM polling thread using the workspace just when the next build was about to start. I'll write a test case to examine this behavior. I assume the previous behavior of blocking the build until the polling is over is desirable, right? At least unless the "do concurrent build" option is checked.

          Reply from Jesse:

          kohsuke@dev.java.net wrote:
          > https://hudson.dev.java.net/issues/show_bug.cgi?id=4202
          >
          > I assume the previous behavior of blocking the build until the polling is over
          > is desirable, right?

          I guess it would be necessary. Not sure, I just was not expecting and did not
          want another (enormous) workspace created.

          > At least unless the "do concurrent build" option is checked.

          Which it was not.

          Kohsuke Kawaguchi added a comment - Reply from Jesse: kohsuke@dev.java.net wrote: > https://hudson.dev.java.net/issues/show_bug.cgi?id=4202 > > I assume the previous behavior of blocking the build until the polling is over > is desirable, right? I guess it would be necessary. Not sure, I just was not expecting and did not want another (enormous) workspace created. > At least unless the "do concurrent build" option is checked. Which it was not.

          I fixed this in 1.320 so that unless concurrent build option is specified, it
          maintains the past behavior of one workspace per job per slave.

          Kohsuke Kawaguchi added a comment - I fixed this in 1.320 so that unless concurrent build option is specified, it maintains the past behavior of one workspace per job per slave.

          Code changed in hudson
          User: : kohsuke
          Path:
          trunk/hudson/main/core/src/main/java/hudson/matrix/MatrixRun.java
          trunk/hudson/main/core/src/main/java/hudson/model/AbstractBuild.java
          trunk/hudson/main/core/src/main/java/hudson/model/AbstractProject.java
          trunk/hudson/main/core/src/main/java/hudson/model/FreeStyleBuild.java
          trunk/hudson/main/core/src/main/java/hudson/slaves/WorkspaceList.java
          trunk/hudson/main/maven-plugin/src/main/java/hudson/maven/MavenBuild.java
          trunk/hudson/main/test/src/test/java/hudson/model/AbstractProjectTest.java
          trunk/www/changelog.html
          http://fisheye4.cenqua.com/changelog/hudson/?cs=20630
          Log:
          [FIXED JENKINS-4202] in 1.320. Test case confirmed my hypothesis.

          SCM/JIRA link daemon added a comment - Code changed in hudson User: : kohsuke Path: trunk/hudson/main/core/src/main/java/hudson/matrix/MatrixRun.java trunk/hudson/main/core/src/main/java/hudson/model/AbstractBuild.java trunk/hudson/main/core/src/main/java/hudson/model/AbstractProject.java trunk/hudson/main/core/src/main/java/hudson/model/FreeStyleBuild.java trunk/hudson/main/core/src/main/java/hudson/slaves/WorkspaceList.java trunk/hudson/main/maven-plugin/src/main/java/hudson/maven/MavenBuild.java trunk/hudson/main/test/src/test/java/hudson/model/AbstractProjectTest.java trunk/www/changelog.html http://fisheye4.cenqua.com/changelog/hudson/?cs=20630 Log: [FIXED JENKINS-4202] in 1.320. Test case confirmed my hypothesis.

            Unassigned Unassigned
            jglick Jesse Glick
            Votes:
            0 Vote for this issue
            Watchers:
            0 Start watching this issue

              Created:
              Updated:
              Resolved: