[Beowulf] openMosix ending

[Robert G. Brown]

On Tue, 17 Jul 2007, Tony Travis wrote:

> Do you really think that Condor is an alternative to openMosix?
>
> I don't know much about Condor, but I thought is was a DRM (Distributed 
> Resource Manager) like SGE. Is it more than that?

I don't know for sure how functional it is as a OM alternative in
reality, but according to its (copious!) documentation, it should be
pretty good.  It basically appears to be a mix of scheduler, DRM, and
migration/checkpointing tool, with the latter fairly recent.  For an
application to migrate, apparently you have to link it to a
condor-supplied checkpointing library (but little else).  I'm guessing
that this library "wraps" all the relevant systems calls so that they
can be sent a signal, save their state information, and encapsulate into
a migration package that actually moves (much as doubtless happens with
Mosix, but in userspace instead of kernel space, which is actually
good).

Now that Condor has a better license (basically an open source license,
although their primary distribution site still has a lot of nonsense on
it) I think it is worth giving a try.  Rumor has it that a lot of groups
running grid-like computers favor it for resource control etc.  This is
indeed the EP application crowd.

I keep planning to try it myself one day soon, but I'm pretty busy with
ongoing projects and just don't have time.  There is a condor VM you can
download for free from VWware -- that and the player might let you
experiment some with minimal personal time lost building software an so
on.

    rgb

>
> The great thing about openMosix is that most 'ordinary' programs migrate. 
> I've thought about using openSSI previously: What's your opinion about that 
> for 'embarrassingly' parallel computation?
>
> Best wishes,
>
> 	Tony.
>

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu