[asterisk-users] Multiple instances of Asterisk on the same host...

[Gordon Henderson]

On Tue, 23 Feb 2010, Tzafrir Cohen wrote:

> On Tue, Feb 23, 2010 at 12:19:31PM +0000, Gordon Henderson wrote:
>> On Tue, 23 Feb 2010, Tzafrir Cohen wrote:
>
>>> But then again, lxc uses much of the work on containers done also by and
>>> for OpenVZ. Sort of like the VMWare/Xen/KVM story all over again, with
>>> lxc playing the role of KVM.
>>
>> And LXC got into the kernel before the others - what that means is anyones
>> guess - probably because it was sponsored/written by IBM?
>
> KVM was not sponsored by any big-name company (RedHat, IBM and the
> others had their bets on Xen, at the time).

Indeed, so it just goes to show..

LXC is growing on me - I like that I can see the underlying 'container' 
root directory from the host - makes it trivial to distribute config 
files, updates, etc. Also takes up minimal disk space as each container 
can be hard links to a 'base' container with the exception of a file files 
in /var/ that get written to. Update the base container and all the others 
update automatically...

Asterisk seems to run OK too, as does dahdi_dummy.. Just ran up 4 
containers and 4 dahdi_tests, one in each container and they're all 
reporting the same (99.997, etc.) (And this is on my test box with 256MB 
of RAM in it)

Will place some test calls through it later on. I reckon I can get about 8 
contaners going on this box... A crude plan is to get container 0 to call 
c1, then to call c2, then ... and c7 does echo() and then point sipp at it 
and see what happens - if it can survive a number of calls on that 
hardware, (1.8Hz Celeron) I'll be more than happy on a proper server...

One thing that does fail is asterisk -p - well, it doesn't fail, just 
prints "Unable to set high priority", so that's possibly something that 
the container isn't allowing - more research required.

Gordon