Please try to change log level to DEBUG, and see heartbeat for ShellBolt is updated correctly.

In ShellBolt, unless subprocess takes numerous time to handle a tuple, heartbeat timeout should not happen. If you are not able to fix the heartbeat issue, you can set topology.subprocess.timeout.secs to quite large value so that heartbeat of ShellBolt never timed out.

Jungtaek Lim (HeartSaVioR)

2017년 2월 24일 (금) 오전 12:45, Kunal Gupta <>님이 작성:
Can you put your code in try and except. 

Create a logger and call log. info within the function of bolt. 

On 23 Feb 2017 21:10, "Mostafa Gomaa" <> wrote:
The code in the bolt itself is not crashing, any ideas?

On Feb 23, 2017 5:35 PM, "Kunal Gupta" <> wrote:
Have you enabled a logs of your PyBolt. py? 

I think your code of bolt is getting crashed.

On 23 Feb 2017 20:59, "Mostafa Gomaa" <> wrote:
Hello all,

I am trying out adding a python bolt in my topology, but it keeps
crashing quite frequently. Before adding this python bolt, my topology
was EXTREMELY stable and would run for months without any issues.
However, after adding the python bolt, the following exception happens
quite frequently:

2017-02-23 15:14:04.080 o.a.s.t.ShellBolt [ERROR] Halting process:
ShellBolt died. Command: [python,], ProcessInfo pid:31162,
name:PythonBolt exitCode:-1, errorString:
java.lang.RuntimeException: subprocess heartbeat timeout
at org.apache.storm.task.ShellBolt$
at java.util.concurrent.Executors$
at java.util.concurrent.FutureTask.runAndReset(
at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(
at java.util.concurrent.ScheduledThreadPoolExecutor$
at java.util.concurrent.ThreadPoolExecutor.runWorker(
at java.util.concurrent.ThreadPoolExecutor$
at [?:1.8.0_121]
2017-02-23 15:14:04.092 o.a.s.d.executor [ERROR]

I am using storm 1.0.2. Any tips or ideas?