[PIG-3853] Pig does not kill all the jobs during shutdown for hadoop2

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Assignee: Aniket Namadeo Mokashi
• Reporter: Aniket Namadeo Mokashi
• Created: 28/Mar/14 21:03
• Updated: 13/May/14 18:35

Description

Current code -

for (Job job : jc.getRunningJobs()) {
                    RunningJob runningJob = getJob(job.getJobClient(), job.getAssignedJobID());
                    if (runningJob!=null) {
                        runningJob.killJob();
                        log.info("Job " + job.getAssignedJobID() + " killed");
                    }
                }

is not sufficient for killing all the submitted jobs. It does not kill the jobs that are submitted to the cluster but are in UNDEFINED state (with appid but no app-master).

Issue Links

• relates to: PIG-3913 Pig should use job's jobClient wherever possible (fixes local mode counters)