[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: woody CDs...

On Thu, 2002-04-25 at 19:30, Richard Atterer wrote:
> On Thu, Apr 25, 2002 at 09:00:50PM +1000, Anthony Towns wrote:
> > What's going to be made available where for release? Just jigdo
> > stuff?
> ISTR Phil wanted to do all isos for i386 and 1..2 CDs' isos for the
> other arches, the rest with jigdo. (Correct me if I'm wrong.)

I was thinking that initially I'd just have the jigdo images available,
so that the mirrors can grab them quickly, and then generate (or perhaps
simply publish) the ISO images that seem worth having.

My first guess at which ones were worth having was, a full set of
source, a full set of i386, and CD1 and/or CD1_NONUS for all the rest,
so that it's possible to bootstrap yourself on any architecture to the
point that you can get jigdo to build the rest.

> > Are the jigdo images for i386 likely to be broken now since some of
> > the packages that were in the archive on the 8th of April will have
> > been removed from the mirrors by now?
> They are broken, I think. For the woody release, a snapshot will be
> taken of the archive at the time of the release. Later, that snapshot
> can be used by jigdo as a fallback when any file changes.

I've got a half written perl script for doing this, but I want it to
pick the files out of the link farm, rather than the main archive (as it
does now) so that it will be immune from changes in the main archive. 
That in combination with the fallback server in the jigdo files (which
I'm going to add in the same script) should make the jigdo files bullet

> > How long's it take to build them all?
> I've never done it myself, but I think it's something like 10..15
> hours including everything (on cdimage, which is usually rather
> loaded).

It took over 2 days for the last full attempt.

> Well, I hope jigdo will be up to it. It certainly works nicely as long
> as no file in the archive changes or disappears...
> [OTOH, with the current 2.2r6 release, for some bizarre reason the
> gentoo source deb has changed its md5sum since the images were
> generated, so source-2 can no longer be regenerated. If we rely on
> jigdo exclusively, such things might happen every now and then.]

I think that must have been due to a single bit error in the md5sum
phase of jigdo file generation on open, which might point towards dodgy
RAM, which might explain open's continued unreliability.  I'll be
checking the jigdo files for validity before signing them off, so this
should not happen again.

I'm testing some new ram at the moment, with a view to swapping the RAM
in open (again) to see if that will help.

Cheers, Phil.
Say no to software patents!  http://petition.eurolinux.org/

|)|  Philip Hands [+44 (0)20 8530 9560]    http://www.hands.com/
|-|  HANDS.COM Ltd.                    http://www.uk.debian.org/
|(|  10 Onslow Gardens, South Woodford, London  E18 1NE  ENGLAND

Attachment: signature.asc
Description: This is a digitally signed message part

Reply to: