Re: [gentoo-user] QEMU/distcc combination question.

[Neil Bothwick]

On Fri, 1 Jan 2016 22:11:34 -0500, waltd...@waltdnes.org wrote:

>   I'm trying to run a distccserver in a 32-bit VM on a 64-bit host, for
> the benefit of my ancient 32-bit-only netbook.  Yeah, "it'll work" using
> the native 64-bit host OS.  But any stuff that links against 32-bit
> libraries is going to be sent back to the netbook to compile locally.
> That defeats the whole purpose of distcc.  This is why I want the 32-bit
> VM to compile for the 32-bit Atom.  Here's the launch script for the
> 32-bit VM on the i3 machine...

I used to take a different approach. Instead of a VM I used a chroot
that was a clone of the netbook, except that make.conf in the chroot
included buildpkg in FEATURES and the netbook's make.conf have --usepkg in
DEFAULT_OPTs. PKGDIR was an NFS share accessible to both.

Then I would simply emerge whatever I wanted in the chroot, then emerge it
on the netbook. No messing around with distcc, which will always run some
stuff on the local system, instead everything but unpacking the package
tarballs was done in the VM.

This approach meant that I could easily script the build process for
several systems, including a 486 box was running at the time. I later
switched to using systemd containers instead of chroots.


-- 
Neil Bothwick

Is it true that cannibals don't eat clowns because they taste funny?


pgpbgFNsCh0se.pgp
Description: OpenPGP digital signature

[gentoo-user] QEMU/distcc combination question.

[waltdnes]

  First, basic definitions.

  distcc ==> compile on machine A for machine B.  No mention of whether
 the processors are different architectures.

  cross-compile ==> compile under architecture A for architecture B

  I'm trying to run a distccserver in a 32-bit VM on a 64-bit host, for
the benefit of my ancient 32-bit-only netbook.  Yeah, "it'll work" using
the native 64-bit host OS.  But any stuff that links against 32-bit
libraries is going to be sent back to the netbook to compile locally.
That defeats the whole purpose of distcc.  This is why I want the 32-bit
VM to compile for the 32-bit Atom.  Here's the launch script for the
32-bit VM on the i3 machine...

#!/bin/bash
qemu-system-i386 -enable-kvm \
   -cpu host -display gtk \
   -drive file=gentoo32.img,format=raw \
   -drive file=linuxswap.img,format=raw \
   -net nic,model=virtio \
   -rtc base=localtime,clock=host \
   -net 
user,hostname=gentoovm,hostfwd=tcp::2022-:22,hostfwd=tcp::3632-:3632,hostfwd=udp::3632-:3632
 \
   -m 3G -name "Gentoo VM" \
   -parallel none \
   ${@}

  Amongst other things, it forwards port 3632 destined for the host, to
the VM.  "ssh -p 2022 root@192.168.123.249" works from the Atom to the
VM.  I get an ssh teminal.  There's no iptables blocking things, and
distccd is running on the VM (it's in the default rc-update level).  Yet
the Atom spits out stuff like...

distcc[8248] (dcc_pump_sendfile) ERROR: sendfile failed: Broken pipe
distcc[8248] (dcc_writex) ERROR: failed to write: Broken pipe
distcc[8248] Warning: failed to distribute shell.c to 
192.168.123.249/6,lzo,cpp, running locally instead

...when trying to outsource the compile.  What am I doing wrong?

-- 
Walter Dnes 
I don't run "desktop environments"; I run useful applications