On Thu, 2009-04-09 at 10:21 -0400, Lamar Owen wrote:
> On Wednesday 08 April 2009 17:36:18 R P Herrold wrote:
> > I would be thrilled to have a simultaneous coordinated
> > release, but the 'leak' of 'patched' torrent instances, and at
> > least two mirrors opening the full ISO set before the
> > coordinated bit flip date and time, leave rather a bad outlook
> > to me as to the ability to make things better through such an
> > 'inverted as to demand' approach
>
> > My $0.02 .... I'd love to be shown a path to avoid the
> > problems on the 5.3 roll-out
>
> Russ, Karnbir, et al:
[...]
> Now, as to the technical issues, it seems to me that a fully ACID compliant
> transactional repository mirror system is possibly one way to eliminate most
> of these issues. Such a system to my knowledge does not yet exist;
If you want to have a technical talk, you probably want to speak to
Matt Domsch who does the mirroring for Fedora (releases 2x a year and
rawhide :).
As of "recent" versions of yum, creating repos. with
--unique-md-filenames means an atomic move from a fully working old DB
to a new one can¹ be done with rename("new-repomd.xml", "repomd.xml").
Which, excepting ext4 ;), is pretty atomic. However my understanding,
from speaking to Matt, is that "release day" problems are almost
entirely due to getting that very last change to all the mirrors in a
timely manner.
Also metalink data can be used to automatically filter out mirrors with
old repomd.xml.
However those both require some sacrifices for people using the GA
version of yum in CentOS-5, so I'm doubtful that's it's worth the pain.
[...]
> But the current pull updating structure doesn't lend itself
> readily to this.
Trying not to speak too much for Matt, but my understanding is that
almost no external mirrors will accept pull mirroring.
¹ My understanding is that no mirrors both with this level of atomic
update, but it's possible.
--
James Antill - james at fedoraproject.org
"I'd just like to see a realistic approach to updates via
packages." -- Les Mikesell