Re: src snapshot archives?

On Sun 20 Sep 2009 at 12:51:01 AM -0400, Thor Lancelot Simon wrote:
>On Fri, Sep 18, 2009 at 02:50:15PM -0700, George Georgalis wrote:
>> 
>> A key reason I began using netbsd was the stability and consistancy.
>> It's never good when validated disaster recovery QA scripts fail.
>
>Let me get this right: your "disaster recovery QA script" mirrors the
>source sets for the daily autobuilds of NetBSD?

no, the qa disaster recovery script responcible for installing the
os pulls netbsd-5 to build a release kernel/world. I don't use it
to get auto builds, but to accelerate the cvs checkout of src.

I used the suite to restore IT services on new hardware after a
key component failed, it was a complex environment, the fact that
the recovery process was 95%+ scripted meant the staff could get
back to work quickly and the process and systems did not need to
be documented/tested. All the build scripts and configuration
management just worked. It was simply a matter of re executing the
validated computer environment and restoring backups files.

>That seems a little odd for me, and very much not a use case anyone
>involved with running the service in question with NetBSD expected or
>intended to support.

simply requesting that a heads up be made if distribution method
is changed.

heck the reason I began using netbsd was documenting and
automating a debian release install was such a PITA. The 'clean'
aspect of netbsd makes it much better for a deployment which
requires assurance that it is as documented. Best way for that is
scripting the deployment and configuration management, and signing
off those files.


What gives? all the files went from a netbsd ftpd to an http
daemon that I've not heard of and all the mirrors now have a goto
nyc host readme. (which gave me errors when I tried to download
files) Yeah, it broke my scripts but the unanswered question is
why the change? and why no announcement? We are talking about the
OS distribution here....(???), I mean was there a problem?

-George