This is a production node on real hardware. I like the strace idea, do you have a workable command line for that?

On Wed, Oct 12, 2011 at 1:13 PM, Mohit Anchlia <mohitanchlia@gmail.com> wrote:
Yes. If you have exhausted all the options I think it will be good to
see if this issue persists accross other nodes after you decommission
that node.

If this is not production and issue is reproducible easily you can
also try using strace with fork option to see if it gets killed at the
same place.

Are these vms?