Uploaded image for project: 'Jenkins'
  1. Jenkins
  2. JENKINS-14087

ScriptTrigger plugin timer issue

    XMLWordPrintable

Details

    • Bug
    • Status: Reopened (View Workflow)
    • Major
    • Resolution: Unresolved
    • scripttrigger-plugin
    • None
    • CentOS, jenkins ver 1.424.6

    Description

      We have noticed some intermittent issues with the ScriptTrigger plugin build timer.The plugin version is 0.11.1.
      There are two issues:
      1. Sometime the build timer does not work per schedule,
      e.g., a job was run more frequently than the scheduled run time.
      2. Sometime the build timer does not work at all,
      e.g., there are no builds kicked off when a build supposedly to be kicked off per schedule.

      Attachments

        Activity

          Is it for ScriptTrigger plugin?

          gbois Gregory Boissinot added a comment - Is it for ScriptTrigger plugin?
          amargono amargono added a comment -

          sorry, yes it is the ScriptTrigger plugin.

          amargono amargono added a comment - sorry, yes it is the ScriptTrigger plugin.
          amargono amargono added a comment -

          Additional Info:
          In the jenkins UI, some jobs only shows "Polling has not run yet."

          Those jobs usually run scriptTrigger every 2-4 hours. And many runs at the same time based on timestamp. The timestamp shows the scriptTrigger-polling.log is created but can not be started because there is no content. So it looks timer trigger to start (at least log file is touched) but nothing is able to run. Is that due to too many scriptTrigger so exceed some thread limitation?

          We have tried to install ver 0.19 and 0.20 but it didn't work either.
          We notice that 0.13 fixed the open files issue and we will give it a try.

          Thanks,
          amargono

          amargono amargono added a comment - Additional Info: In the jenkins UI, some jobs only shows "Polling has not run yet." Those jobs usually run scriptTrigger every 2-4 hours. And many runs at the same time based on timestamp. The timestamp shows the scriptTrigger-polling.log is created but can not be started because there is no content. So it looks timer trigger to start (at least log file is touched) but nothing is able to run. Is that due to too many scriptTrigger so exceed some thread limitation? We have tried to install ver 0.19 and 0.20 but it didn't work either. We notice that 0.13 fixed the open files issue and we will give it a try. Thanks, amargono
          yanqin1 Yan Qin added a comment -

          This kind of issue is kind of related with scalable issue. When we have only a few projects using scriptTrigger, it seems ok. At present we have 125 jobs using ScriptTrigger so the hanging issue happen more frequently. Each ScriptTrigger plugin calls one script which takes less than 1 minute.

          yanqin1 Yan Qin added a comment - This kind of issue is kind of related with scalable issue. When we have only a few projects using scriptTrigger, it seems ok. At present we have 125 jobs using ScriptTrigger so the hanging issue happen more frequently. Each ScriptTrigger plugin calls one script which takes less than 1 minute.

          Thanks everybody for trying to use the scripttrigger Jenkins plugin.

          @amargono
          Do you know why there is the log 'Polling has not run yet'?
          Are all your slaves off?

          @Yan Qin
          Polling is always resources consuming.
          125 jobs polling resources at the same time is a limit.I suggest you should split your instance in multiple Jenkins instances.

          gbois Gregory Boissinot added a comment - Thanks everybody for trying to use the scripttrigger Jenkins plugin. @amargono Do you know why there is the log 'Polling has not run yet'? Are all your slaves off? @Yan Qin Polling is always resources consuming. 125 jobs polling resources at the same time is a limit.I suggest you should split your instance in multiple Jenkins instances.
          yanqin1 Yan Qin added a comment -

          Hi, Greg:

          Thank you for your quick response.

          These 125 job scripttrigger do not happen at the same time. And each trigger takes within 1 minute. That means the probability of having many triggers at the same time is low. I checked all timestamp of log when trigger running normally. At most 3 job triggers happen at the same time. So that shall not be issue.
          But when I checked the timestamp of log when scripttrigger did not work, the timestamp of more jobs are same (maybe more than 3 but I do not remember very clearly).

          some of jobs use the same timer format
          0 */3 * * *, 40 */2 * * *
          Will that be problem?

          Lastly, about your question to allan, slave is not off. In the jenkins web ui "scriptTrigger" only shows "polling is not run yet".And the scriptTrigger.log has no content although the timestamp shows it did recreate on time. So it does not looks like timer issue. It seems the log file is touched but can not proceed. Do not know where it is blocked.

          yanqin1 Yan Qin added a comment - Hi, Greg: Thank you for your quick response. These 125 job scripttrigger do not happen at the same time. And each trigger takes within 1 minute. That means the probability of having many triggers at the same time is low. I checked all timestamp of log when trigger running normally. At most 3 job triggers happen at the same time. So that shall not be issue. But when I checked the timestamp of log when scripttrigger did not work, the timestamp of more jobs are same (maybe more than 3 but I do not remember very clearly). some of jobs use the same timer format 0 */3 * * *, 40 */2 * * * Will that be problem? Lastly, about your question to allan, slave is not off. In the jenkins web ui "scriptTrigger" only shows "polling is not run yet".And the scriptTrigger.log has no content although the timestamp shows it did recreate on time. So it does not looks like timer issue. It seems the log file is touched but can not proceed. Do not know where it is blocked.
          wael Wael Darwich added a comment -

          I used this plugin for long time, and when I had to move to a different server and create new Jenkins system and reinstall all plugins I started getting this error, now I only get for my job "Polling has not run yet." and I get it always!

          I am using Jenkins ver. 1.517 and scripttrigger 0.28

          wael Wael Darwich added a comment - I used this plugin for long time, and when I had to move to a different server and create new Jenkins system and reinstall all plugins I started getting this error, now I only get for my job "Polling has not run yet." and I get it always! I am using Jenkins ver. 1.517 and scripttrigger 0.28
          wael Wael Darwich added a comment -

          I've been checking this for awhile, and noticed that in rare case polling starts on time, but it does not use the current script in the job configuration, it uses a previous version of it!! the script is a simple echo message, when I changed the script to echo a different message it was using an old one!!!

          wael Wael Darwich added a comment - I've been checking this for awhile, and noticed that in rare case polling starts on time, but it does not use the current script in the job configuration, it uses a previous version of it!! the script is a simple echo message, when I changed the script to echo a different message it was using an old one!!!
          gbois Gregory Boissinot added a comment - - edited

          Unfortunately, I can't reproduce it.
          It is OK in my environment.

          @wael for your previous issues, please could you test from ScriptTrigger 0.30.

          gbois Gregory Boissinot added a comment - - edited Unfortunately, I can't reproduce it. It is OK in my environment. @wael for your previous issues, please could you test from ScriptTrigger 0.30.

          In our environment we experienced the same issue. It happens when the process which the trigger was running freezes.
          In our case the Shell script from the trigger runs a Java tool which freezes and the triggers for all jobs stop.
          When the process is killed, the triggers start running again.

          raykovl Lyubomir Raykov added a comment - In our environment we experienced the same issue. It happens when the process which the trigger was running freezes. In our case the Shell script from the trigger runs a Java tool which freezes and the triggers for all jobs stop. When the process is killed, the triggers start running again.
          xgwang John Wang added a comment - - edited

          Gregory , i have got the same issue as Wael on ScriptTrigger 0.30. (the issue is as same as 2th issue of description)
          MY jenkins version is 1.6.05

          IN addition , could you have a workaround for that issue?

          xgwang John Wang added a comment - - edited Gregory , i have got the same issue as Wael on ScriptTrigger 0.30. (the issue is as same as 2th issue of description) MY jenkins version is 1.6.05 IN addition , could you have a workaround for that issue?

          People

            Unassigned Unassigned
            amargono amargono
            Votes:
            3 Vote for this issue
            Watchers:
            7 Start watching this issue

            Dates

              Created:
              Updated: