F7-T4 Install Woes (LVM, RAID, grub)

[Will Woods]

On Wed, 2007-05-09 at 08:23 -0500, Brian Hanks wrote:
> --I have BZ'd several software raid and LV issues over the months, but
> they get pretty --much ignored as not reproducible.  See earlier post
> this list about problems "FC7T4 Not --Seeing Pre-existing software
> raid arrays"  Seems that you must go to console 2 prior to --setting
> up your partitions and make sure your raid arrays and LVs are there
> and active.  --Then allow anaconda to move to the partitioning phase
> and use custom configuration.  I --first noticed problems with raid
> arrays and LVs during FC6Test days and they continue --with the change
> to using the SCSI interface for all disk detection and enumeration.
> 
> I hear you, they do seem to get ignored. 

Seriously, we're not ignoring them - I personally have spent a *lot* of
time trying to reproduce these problems. But either I lack the needed
hardware (e.g. Sil 3112 RAID) or I don't have enough information about
the drive/partition layout, and I can't trigger the problems.

>  The truly frustrating part to me is that features that were working
> in RH9 (and before) are now showing up as broken, and they don't seem
> to get much attention.  I admit that I don't have as much time to
> devote as I did in the old days, but I still try to do my part.  Seems
> to me that we need to focus a little more effort on regression
> testing.  

It's frustrating for us too - we really *want* to fix this stuff, we
spend hours trying to reproduce it, but in the end we lack some key
piece of info or hardware to make it happen.

Where are your bug reports? Give us some BZ numbers, and we'll do the
best we can to get them on the blocker list and get 'em fixed for F7.
But we can't fix your problems without a lot of help from you - after
all, we don't have your hardware and we don't have your disk layout.

> Progress isn't progress unless all of the pre-existing features still
> work.  I face this every day in my work, and we spend a significant
> amount of time regression testing to ensure that our releases don't
> impact the customer. 

Are you suggesting I spend an *insignificant* amount of time doing
regression testing? We do the best we can, and with a 6-month release
cycle the next one's always just around the corner. It's a monstrous
task and it couldn't happen without your active participation.

So thanks for continuing to help, despite your frustration. Please keep
testing and try to be patient. The more we all work together the faster
these things will get done.

-w
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.fedoraproject.org/pipermail/test/attachments/20070509/b5d76eb0/attachment.bin