On Tuesday 10 May 2005 12:18, Derick Swanepoel wrote:
On 5/10/05, Frank Mehnert fm3@os.inf.tu-dresden.de wrote:
On Tuesday 10 May 2005 11:01, Derick Swanepoel wrote:
I'm trying to start three VMs from grub (one master and two slaves), but the tftp provider fails to load the third kernel. This file is the same as the second VM's kernel, which loads successfully.
Hmm, no idea. Has your L4Linux a driver for your network card included (it should not). Please post the full log file.
The kernel used for the master VM has a driver for my network card, but the two slave kernels do not. Note that everything works fine if I start one master and one slave, or two slaves, but not one master and two slaves.
However, sometimes it does work, but very infrequently.
This is exactly what I mean. If your master L4Linux instance grabs the network card, the tftp task owns the network card no longer and therefore is not able to load something. Note that tftp and L4Linux access the same hardware device. This works as long as they do it sequentially.
To overcome this problem, you could load the slave L4Linux instances with grub, pass them to the bmodfs server and then start them using the bmodfs server (see generic_fprov/examples/bmodfs).
I've attached the boot log (it's very long and somewhat hard to make sense of because the messages of the master and slave kernel are interleaved :( )
Hmm. According that log it seems that the master L4Linux instance initialized the netword card just while loading the seconds slave instance, eth0 of Linux is up immediately after the tftp failed.
The right solution for this is to use only _one_ driver for your network device and to share the driver between tftp and L4Linux. Both tftp and L4Linux should also support the oshkosh backend.
Or you don't use tftp at all and load the binaries using bmodfs (maybe this is easier to setup).
Frank