[rb-general] [FOSDEM16] Reproducible FreeBSD and variants

[Holger Levsen]

Hi,

On Freitag, 12. Februar 2016, Steven Chamberlain wrote:
> It is there:
> http://lists.reproducible-builds.org/pipermail/rb-general/2016-February/000
> 031.html but, the mailing list scrubbed my GPG MIME signature, and I think
> when it did that it broke the DKIM signature also :(

how strange. Again I only received it once…
 
> I looked at some other options, and really like the idea of using
> (older) hardware I already have at home for this.  With 'free' hardware
> + the cost of electricity, it is still cheaper than server rentals or
> clouds I think.  I'd need to utilize it really efficiently though.

wow. where do you live that electricity is that cheap? :)
 
> Helmut's rebootstrap jobs for kfreebsd->linux-armhf and linux-mipsel
> completed yesterday for the first time (yay!).  I'll enable diffoscope
> for the next run.

yay!
 
> pbuilder doesn't work yet on kfreebsd (patch in BTS for 4+ years but
> never applied!).

well, pbuilder got a new maintainer recently… ;)

> It was quite bad on kfreebsd, taking several minutes for schroot to
> unpack and install a few build-deps.  I definitely need to work on it.

tmpfs makes all the difference.
 
> > So currently we're building 60*24/8*32=5760 packages a day (/8 because
> > average build time is 8min and *32 because we have 32 amd64 builders).
> Thanks for providing these numbers;  they'll be useful for me to
> benchmark against.
> 
> > I suspect the numbers will look similar for armhf, but I'll leave that as
> > an excercize for the reader :)
> 
> I was curious about those too, if there are any ways to speed up the
> armhf builds (and mipsel if we have hardware for that someday).  But,
> one thing at a time...

well, something more like 60*24/35*42 amd then again if you assume the fastest 
build is 5min, you get 60*24/30*42… what would speed up testing armel the 
most, would be more ram and enabling building in tmpfs.

you can optimize as much as you want, if your disk is a disk and not a 
ramdisk, you will loose against a ramdisk.


cheers,	
	Holger
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 828 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.reproducible-builds.org/pipermail/rb-general/attachments/20160212/565d879c/attachment.sig>