remote agent crashing when being farmed

[Nhut Lam]

Hi,

I'm getting the following error when I farm off a remote agent to run a bamboo plan.

I'm not getting an error when the remote agent is dedicated to running the plan.

It seems to start, and then after say 30 minutes the remote agent will crash with this error.

We need to run parallel remote agents because our plans are taking a 3+ hours to run.

does anyone know how to prevent the Bamboo agent from crashing when two or more agents are running the same plan?

Log :

ERROR | wrapper | 2012/06/12 17:43:26 | JVM appears hung: Timed out waiting for signal from JVM.
ERROR | wrapper | 2012/06/12 17:43:29 | JVM did not exit on request, terminated
STATUS | wrapper | 2012/06/12 17:43:31 | JVM exited in response to signal SIGKILL (9).
ERROR | wrapper | 2012/06/12 17:43:31 | Unable to start a JVM
STATUS | wrapper | 2012/06/12 17:43:31 | <-- Wrapper Stopped

[Nhut Lam]

problem with the agent. Please see above.

[Marcin Cabaj]

Increase your wrapper.ping.timeout (default=30) and wrapper.ping.interval (default=5) values in conf/wrapper.conf, e.g:

wrapper.ping.timeout=900
wrapper.ping.interval=30

more info: http://wrapper.tanukisoftware.com/doc/english/prop-ping-timeout.html

[Nhut Lam]

Thanks for your help.

We resolving why the timeout was occuring.

The agent being farmed to didn't have enough memory, and couldn't respond on occaision to the heart-beat. After increasing the memory from 500 MB to 2GB, everything is fine.

[Marcin Cabaj]

Increase your wrapper.ping.timeout and wrapper.ping.interval values in conf/wrapper.conf, e.g:
* wrapper.ping.timeout=900
* wrapper.ping.interval=30

more info: http://wrapper.tanukisoftware.com/doc/english/prop-ping-timeout.html

[Przemek Bruski]

What do you mean by "farmed"?

Which JDK are you using to start the agent?

[Nhut Lam]

Hi,

By farmed I mean executing a Bamboo plan (which has multiple jobs) with 2 or more bamboo agents.

The java version we are using is : 1.6.0_30

when i run the plan with a single agent it runs and completes fine. With multiple agents executing jobs in parallel I get the above error.