Multiple Mac OS 9.2 running /almost/ simultaneously

Samuel Rydh mol-general@lists.maconlinux.org
Wed, 19 Mar 2003 14:08:22 +0100


> Hi all there, maconlinux friends ;)
> 
> First, I must apologize for my poor english, it is not my mother tongue.
> 
> I have run four MacOS 9.2 instances under maconlinux 
> 0.9.68. Inside each work one QuarkXpress with TCP/IP XTension
> from Techno-Design. All work fine, each Mac have own IP in our
> LAN and running from within TCP/IP XT is fully accessible from outside. 
> Each MacOS instance start from own MacOS partition. I have only one
> problem - when our web application send a job to TCP/IP XT on each
> instance of MacOS, maconlinux have behavior as a stack. Processing 
> of a job in each Quark reach to about 50%, then stop, next MacOS 
> instance also just like first, then third and then fourth, who
> however execute task to the end. Now, in reverse order his
> task complete third, second and first MacOS instances.
> 
> How I can be able to make different instances of
> MacOS to do this work /almost/ simultaneously?

Well, it should be done almost simultaneously. The mol instances
are scheduled as any other linux process (each of them
should have about a 10ms time-slice).

If you are talking about a scheduling behavior involving
longer time-scales than that, then you might be seeing problems
caused by the network setup. I'm not confident that
the sheep_net driver can handle multiple mol sessions
(it sounds like you are using it). The tun driver probably
works better. However, you will have to use iptables
to setup proper IP-forwarding.

/Samuel