up vote 0 down vote favorite
share [g+] share [fb]

A simple jruby process is freezing our Amazon EC2 Cent OS 5.4 instance. To get it back we have to reboot the complete VM from Amazon Mgmt Console. 

Is there way to prevent an app from freezing the OS?

How do we find out what caused the OS to become unresponse?

Will /var/log folder contain any useful log?. Thanks.

Update 17 Dec 2011: We found this was happening because of 2048m memory allocated to a jruby process running a rake task - delayed_job. Reduced the memory to 512m to fix the issue. It appears delayed_job/JRuby with 2048m seems a deadly combo.

link|improve this question
You need to check your logs for possible errors, bugs. A single memory leak is sufficient to bring the system down. It will be helpful if you can collect statistics about your system. – Khaled Dec 10 '11 at 7:56
feedback

1 Answer

up vote 0 down vote accepted

If lowering the memory limit helped with this, you were probably causing the machine to start swapping. Swapping is a surefire way to kill the performance of any machine, and I'd imagine it's even worse on a VM. Swapping can easily kill the machine in a way that it looks like it's completely frozen, because you can't log into it or interact with it in any real way (Well, ping generally works, but that's about it).

link|improve this answer
you hit the nail on the problem. – so_mv Dec 19 '11 at 15:43
feedback

Your Answer

 
or
required, but never shown
discard

Not the answer you're looking for? Browse other questions tagged or ask your own question.