[ale] Private Tomcat 6.0 Server and VMware Server 2.0

[Michael B. Trausch]

On 02/19/2010 06:58 PM, Michael H. Warfield wrote:
> On Fri, 2010-02-19 at 17:23 -0500, Michael B. Trausch wrote:
>> On 02/19/2010 12:43 PM, Michael H. Warfield wrote:
>>> (although OpenVZ as announced development efforts
>>> to support 2.6.32 with a target of about a month from now).
>
>> That's the one thing that has been annoying me lately about OpenVZ.  LXC
>> is pretty good, though there is one piece of software that I have yet to
>> get working underneath it (FreeSWITCH, my phone system).  It will,
>> however, run under OpenVZ.  I'd probably still be using that if they
>> would just track the mainline and keep up with it.  I find myself
>> wondering just why they don't do that.  I guess if they go with 2.6.32,
>> they won't go again for supporting another kernel until 2.6.46.  :-P
>
> Their stated goal has been to first and foremost address the enterprise
> distributions.  For them, "mainline" has been 2.6.18 of late for RHEL
> and CentOS.  Tracking a rapidly developing kernel is difficult and an
> immense amount of work, especially when the rapid development in exactly
> where most of your changes are (container and namespace and
> virtualization).

I can understand that, though I would think that it would be less effort 
to keep up continuously as opposed to working on one version and then 
forward-porting to something 14 major releases later.  I don't actually 
track changes to the kernel closely enough to have any educated guess on 
that, but I know that there are out-of-tree modules that keep up in that 
manner, so I at least I know it's not impossible.

>> Hopefully, I will have some time to figure out what FreeSWITCH's issues
>> are with running under LXC and get them solved, though.  I needed to use
>> a very new kernel so my hand was forced away from OpenVZ/Proxmox.
>
> Yeah, that's curious.  I'd really be interested in knowing where the
> problems are there.  If they are network related, I might have a few
> suggestions (like avoiding macvlan for the moment and stick with bridged
> - there are some corner cases which are not quite working properly).

I am using bridged networking, so I have to assume that there is 
something that isn't namespaced in LXC that is namespaced in OpenVZ that 
is throwing the behavior off.  I just need to get to it... until then, 
my home phone is, well, down.  :-/

	--- Mike

-- 
Michael B. Trausch                                    ☎ (404) 492-6475