hasDuplicateHistory is very inefficient and can make the Jenkins master unresponsive

This issue is archived. You can view it, but you can't modify it. Learn more

XMLWordPrintable

      See this stacktrace

      "SharedConfigurationSynchronizer thread" #9192 daemon prio=5 os_prio=0 tid=0x00007fb33c134000 nid=0x3718 runnable [0x00007fb336bec000]
   java.lang.Thread.State: RUNNABLE
	at java.io.UnixFileSystem.getBooleanAttributes0(Native Method)
	at java.io.UnixFileSystem.getBooleanAttributes(UnixFileSystem.java:242)
	at java.io.File.exists(File.java:819)
	at hudson.plugins.jobConfigHistory.HistoryFileFilter.accept(HistoryFileFilter.java:41)
	at java.io.File.listFiles(File.java:1291)
	at hudson.plugins.jobConfigHistory.FileHistoryDao.getRevisions(FileHistoryDao.java:335)
	at hudson.plugins.jobConfigHistory.FileHistoryDao.getRevisions(FileHistoryDao.java:325)
	at hudson.plugins.jobConfigHistory.FileHistoryDao.getRevisions(FileHistoryDao.java:320)
	at hudson.plugins.jobConfigHistory.FileHistoryDao.hasDuplicateHistory(FileHistoryDao.java:559)
	at hudson.plugins.jobConfigHistory.FileHistoryDao.checkDuplicate(FileHistoryDao.java:581)
	at hudson.plugins.jobConfigHistory.FileHistoryDao.saveItem(FileHistoryDao.java:269)
	at hudson.plugins.jobConfigHistory.JobConfigHistorySaveableListener.onChange(JobConfigHistorySaveableListener.java:54)
	at hudson.model.listeners.SaveableListener.fireOnChange(SaveableListener.java:81)
	at jenkins.model.Jenkins.save(Jenkins.java:3134)
	- locked <0x00000006c8382f38> (a hudson.model.Hudson)
	at hudson.BulkChange.commit(BulkChange.java:98)
	at com.cloudbees.opscenter.context.config.ToolsInstallerSnippetHandler.consume(ToolsInstallerSnippetHandler.java:90)
	- locked <0x00000006c5a42bd8> (a com.cloudbees.opscenter.context.config.ToolsInstallerSnippetHandler)
	at com.cloudbees.opscenter.context.config.ConfigurationSnippetRefresher.execute(ConfigurationSnippetRefresher.java:56)
	at hudson.model.AsyncPeriodicWork$1.run(AsyncPeriodicWork.java:100)
	at java.lang.Thread.run(Thread.java:745)

   Locked ownable synchronizers:
	- None

      Looking at the plugin code, it seems to actually read all existing history files for a config file just to compare the new state to the latest saved state. This should be more efficient as it only cares about the latest one anyway:
      https://github.com/jenkinsci/jobConfigHistory-plugin/blob/b2d39c7a556e77c8c5bdc123b9131362ebd5d52c/src/main/java/hudson/plugins/jobConfigHistory/FileHistoryDao.java#L815

      So it seems that in JCH either you don't de-duplicate, or keep short histories, for now.

            Assignee:
            Félix Belzunce Arcos
            Reporter:
            Félix Belzunce Arcos
            Archiver:
            Jenkins Service Account

              Created:
              Updated:
              Archived: