Re: [9fans] Let's get VM configs onto the Wiki.

[andrew zerger <rhoyerboat@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2760 bytes --]

So many configs- is true, since the configuration of the host itself is
critical to the virtual platform, and your distroVersion/openbox vs.
distro2Version/openbox.

Especially the successful archlinux/gentoo/lfs qemu-kvm guys will be
like----  well first I (already) had a kernel compiled for my hardware
(lspci, kernel.config), then I enabled KVM support in the kernel, and
tun/tap support, then it just worked, because of course, all of that is in
the wikis/docs for those necessary steps.

If anyone has trouble I would recommend the docs say, compile your own
virtual host and glean issue/resolution wiki from what transpires there,
otherwise it will be a distro specific problem on the hardware support side
of distro->vm->(no-longer hardware phase)guest, and that distro/VM team
would be more interested to know how what is broken than anyone looking at
plan9 code.

Or- not being some kind of gentoo snob, if I had/'there were' some docs on
how to get host-side information on how many supported/unsupported
syscalls, etc, plan9 made to qemu, I think that would be useful for
improving the performance of plan9 on virtual hardware, but I'm not sure.
Just letting my mind wander at the end of the day. Those docs on debugging
qemu guests probably exist somewhere I won't see right away.

regards,
andrew



ps,
Here's a really bad startup script for qemu-kvm, haha
(not really, its just really bad .. okay 1 line)

kvm -net nic,macaddr=$DISTMAC \
    -net tap,ifname=$DISTTAP,script=no,downscript=no \
    $DVMOPT \
    -hda $DISIMG -m $MVMRAM -daemonize



On Tue, Nov 22, 2011 at 3:58 PM, erik quanstrom <quanstro@quanstro.net>wrote:

> > What would be *really* helpful is if people who have actual real live
> > running this minute Plan 9 under some VM system would post their
> > *specific* VM and Plan9 configuration files to the Wiki.
> >
> > Several people claim to be running Plan 9 under assorted VMs, but it's
> > very difficult for others to reproduce that success, and every time I ask
> > someone for specific configs the response is "well that was months ago
> and
> > I don't use it any more" or suchlike.
> >
> > Not that I don't believe them, but basically I don't believe them ;-)
>
> i think the problem is that there are so many configurations.
> there are at least
>
>        vm versions * vm config * real hardware
>
> many of them.  hardware passthrough has got to be one of the
> least appealing ideas that's come out of virtualization.  you get
> all the complications of a virtual environment, coupled with the
> convenience and sheer joy of dealing with hardware.
>
> - erik
>
>


-- 
⎼⎺⎺├@┼␊├├≤-␍⎼␊▒␍:/␤⎺└␊/⎼␤⎺#

[-- Attachment #2: Type: text/html, Size: 3484 bytes --]