Moving from "type forward" to "type static-stub"
Oscar Ricardo Silva
osilva at scuff.cc.utexas.edu
Fri Sep 21 15:45:43 UTC 2012
On 09/21/2012 02:55 AM, Adam Tkac wrote:
> On Thu, Sep 20, 2012 at 07:49:08PM -0500, Oscar Ricardo Silva wrote:
>> I have several recursive, caching BIND servers that were running the
>> Redhat package of BIND. Our servers started crashing because of a
>> bug (previously identified AND fixed by ISC) so we've decided to
>> ditch that version and run from source, 9.9.1-P3. (I'm still not
>> sure why redhat decided to use the rc1 version of source on which to
>> build their rpm ... seriously ... the bug that hit us was fixed in
>> rc2 AND the final release)
> Because rc2 was released too late to get it into RHEL 6.3... Btw which is the
> bug that bothers you? Why don't you report it to RH bugzilla?
> Regards, Adam
Case opened with Redhat: 2012-07-02
Bug: 837165 (opened on 2012-07-02)
Proposed patch (by Adam): 2012-07-09
Patch released: 2012-07-23 (but not noted in Bugzilla)
Notification of bug fix: 2012-07-31
Number of crashes after
bug reported: 3 (2 initial crashes, 3 more after)
We did report it and a patch was released AFTER our servers started
crashing on the known bug. Even then, when one server had crashed twice
and we were holding our breath on the others, it took weeks for the
patched version to come through.
I don't want to get in a war of words but here's the first line from the
9.8.2rc1 Release Notes:
BIND 9.8.2rc1 is the first release candidate of BIND 9.8.2.
Not only is it a release candidate but it's the FIRST release candidate.
Also, using the dates on the release notes:
2012-06-21 RHEL 6.3 announced as available
I wonder if there's no process to revisit a package after a final
version has been released. I'm aware that you have to pick something
but if it's not a final version why not go back after the official
version is available.
Sorry for going way off topic (package version vs an actual BIND
problem) but the question was asked. Personally it drives me nuts when
people just complain on a mailing list but then don't report it through
official channels so I wanted to show that we did do that.
More information about the bind-users