[Gllug] Reia

[James Laver]

On Sat, February 21, 2009 12:37 am, John Hearns wrote:
> 2009/2/20 James Laver <gllug at jameslaver.com>:
>> On Fri, February 20, 2009 5:40 pm, - Tethys wrote:
>>> Does each scheduler have a pool of
>>> processes for which it's responsible? Can a process move between
>>> schedulers (and hence cores)?
>>
>> Yes, including moving processes over a network.
>
> A reference to this would be interesting.
> As someone who has looked after a Mosix cluster (in the basement of
> the Natural History Museum) I would like to see what the requirements
> for this are - eg. identical kernels etc.

http://www.erlang.se/euc/08/euc_smp.pdf , section 5.2.1 or ctrl-f for
'migrate'.

Essentially they're not real processes, they're actually greenthreads. The
'kernel' in question is actually the erlang VM. Lots of Telcos (erlang
serves a lot of telecoms*) use this facility to load balance their
systems. Shift the process to another node if your node starts taking too
much work and everything is better.

--James
* Ever had a call dropped? A bug in one erlang process won't affect the
rest of them so it can happily fail in one process and the others be
unaffected. And all this in greenthreads :)
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug