On Mon, 2011-10-24 at 10:57 +0200, Jes Sorensen wrote:
I realize this is last minute, but I just realized this wasn't marked as
a blocker already, and I am still pretty new to the Fedora package process.
I would like to recommend this one being accepted as a blocker for F16:
The bug in question is related to BIOS raid, in particular when
installing root on the raid device. The previous mdadm package has a bug
in it which prevents dirty raids from being assembled at boot, which is
fixed in mdadm-3.2.2-12. This affects both booting from the install
image and installed systems!
The problem is particularly nasty since if one configures a new BIOS
raid, it will be marked dirty straight away and install will be
impossible. In addition, if one has a clean raid and upgrades or
installs to it, but then manages to dirty it, boot will no longer be
possible and one cannot run updates. In other words, a zero-day fix
would be highly suboptimal.
We picked this up and accepted it already, but for the public record,
the process is documented here:
To distil, you just mark the bug as blocking the appropriate tracker
bug, in this case, 'F16Blocker'. There is a long-standing problem in
that you need editbugs privileges to do this (so you need to be either a
packager or a bugzapper), whereas we really want *anyone* with an
account to be able to propose a blocker bug; I've had an issue filed
with RH's Bugzilla team for a while requesting that they change this,
but no joy so far. :(
The fallback if you don't have the privileges to mark a bug as blocking
another is to alert someone on the QA team who does; mailing this list
or the test list will usually achieve that.
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora