Open Stack

Note; I haven't finished debugging the glusterfs job yet.  This
relates to the OOM that started happening on Centos after we moved to
using as much pip-packaging as possible.  glusterfs was still failing
even before this.

On 04/01/2015 07:58 PM, Deepak Shetty wrote:
> 1) So why did this happen on rax VM only, the same (Centos job)on hpcloud
> didn't seem to hit it even when we ran hpcloud VM with 8GB memory.

I am still not entirely certain that hp wasn't masking the issue when
we were accidentally giving hosts 32gb RAM.  We can get back to this
once these changes merge.

> 2) Should this also be sent to centos-devel folks so that they don't
> upgrade/update the pyopenssl in their distro repos until the issue
> is resolved ?

I think let's give the upstream issues a little while to play-out,
then we decide our next steps around use of the library based on that
information.

thanks

-i