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

Re: DRAFT Fedora 9 Installation Test Plan

Jesse Keating wrote:
On Thu, 17 Jan 2008 14:05:12 -0500
James Laska <jlaska redhat com> wrote:

That might be a bit aggressive ... but no better way to tell than
trying it.  A big outstanding piece is to define what tier#1 includes.

How do folks feel about ...

* Install Source / URL
* Install Source / NFS
* Install Source / NFS ISO
* Install Source / DVD
* Package Sets / Default Package Install
* Package Sets / Minimal Package Install
* Partitioning / ext3 on native device 	
* Partitioning / rootfs on LVM device 	
* Partitioning / rootfs on RAID1 (not sure about this one)
* User Interface / Graphical Installation
* User Interface / Text-mode Installation
* User Interface / VNC Installation 	

The intent is that the tier#1 test results provide a 10,000 ft view
of how usable a given tree is.  From there we might dive deeper
involving specific hardware setups and complex environments.

Here is my take.  Installer doesn't really get as much of a benefit of
pre-releases as other software does.  For that reason, installer should
be pretty solid for Alpha, even more solid for Beta.  If things aren't
found at Alpha and fixed for Beta we don't really get a significant
testing of the fix before final and that's a bit hard to consume.

The earlier the better ... the price of late installer code changes is much too high. So here's the tight-rope walk I see right now ...

Trying to get something usable for (current rawhide fun) ... vs ... running a test suite to have a more complete picture of the stability of the software.

Perhaps we keep tier#1 as a measure of how useful a tree is (good for the quick check on rawhide). And extend the alpha release criteria to be more comprehensive?

So I
would concentrate pretty hard on the installer cases for alpha and less
about the content being installed.

"less about the content being installed" ... do you mean package sets or something else?

Also, testing the new installer features would be good, like resizing,
encryption, url method, etc...

Definitely ... there's quite a bit of new features for F9 that could use some attention.



 James Laska         -- jlaska redhat com
 Quality Engineering -- Red Hat, Inc.

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