[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [K12OSN] OT: Hardware question for thin client

SSI clustering is similar to SMP processing.  If you lost a processor in an
SMP box you would lose the processes it was running.  Same thing applies if
you pull the plug on an oM node.  The processes are gone.  (Except in the
one case where I was running pico on a node and pulled the plug and it
still kept right on running on the master system, magic I guess...).  As
far as the entire  "master node" box hanging when a client node drops, I
haven't seen that myself...

At any rate, I'd suggest heading over to the openMosix general list,
they'll be glad to discuss these and any other openMosix related issues
with you I'm sure:


You might want to check out the docs as well, they have put together a lot
of information on the nuts & bolts of how it works and what it can/cannot


James Jensen

--- Dennis Daniels <ddaniels magic fr> wrote:
> Thanks for all the tips here! You've cleared up a question that had been 
> buggin me for a bit... so if I understand correctly... then this might 
> be something doable it's 'just' openmosix doesn't know how to handle a 
> client 'winking out' very well.
> If, and this a big IF, openmosix had better thread switching it would 
> work, having lots of thin clients sharing processes using spare cycles?
> Is it possible through LTSP to keep a node from being shut down through 
> the terminal?...
> am I wandering too far off topic?
> Dennis
> Shawn Austin wrote:
> > The problem with that is if an openmoxis node is rebooted, it will 
> > generally kill the application and hang the machine that it was
> launched 
> > from.  A server cluster makes much more sense for things like this.


Do you Yahoo!?
Yahoo! Calendar - Free online calendar with sync to Outlook(TM).

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]