[CentOS-docs] Updated How to Setup a Software RAID on CentOS 5

Thu Apr 30 17:49:47 UTC 2009
R P Herrold <herrold at centos.org>

On Thu, 30 Apr 2009, pjwelsh wrote:

> On 04/30/2009 10:22 AM, R P Herrold wrote:
>> ...
>> To respond to 'the consensus ... overwhelmingly' remark, the
>> mice also overwhelmingly voted to bell the cat.  Counting
>> noses does not make a bad answer more correct; using raid
>> rather than flat RO /boot partitions is still less robust

> "less robust" !=  "bad answer"  ;-)

I understand your position; I do not concur as it introduces 
failure points in my opinion; we vary

> easy. It lends credibility to the idea that CentOS is more than a
> knock-off RHEL.

I have no such goal to court public opinion, and I think 
neither does the core value of the project;  that people feel 
a need to use CentOS as a locus to contribute is a matter 
beyond the core scope mandate of a strict rebuild project

The CentOS core mandate, to me, is to elide trademarks from an 
upstream FOSS sources rebuild; solving the issue of the 
non-free updater solution upstream; and preserving (lovingly) 
all bugs to match upstream.

I see NO shame in being called knock-off _out of_ a commercial 
product, any more than Red Hat should feel shame in 
stabilising the enormous effort of the free software community 
that preceded them (and that continues independent of, or in 
conjunction with them) _into_ a commercial product.  This is 
the point of FOSS [ESR, and the 'chasing the tail-lights' 

RHT is a 'pure play' FOSS company by and large; NOVL much less 
so; ORCL to my thinking has been a 'white hat' when it decides 
to open something; JAVA [ne SUNW], less so.  See my prior post 
as to Java -- I am encouraged by recent events as to Java.

I think people who are willing to let themselves feel slighted 
into being _just_ a 'knock-off' need to clarify their 
thinking.  It is their issue to solve.  I am not in a 
'credibility lending' business -- I speak with my results, and 
I do not let others triangulate me so simply.

The mandate is the point of my comment to Farkas Levente 
earlier in centos-devel ML today, that there is a slight abi 
change in the gcc across point releases of CentOS (and one 
assumes, in its upstream, Red Hat product) in 5 in a minor and 
slightly tested side package.  It happens; when material a bug 
is filed; it is NOT the end of the world.

I saw the gcc ABI change in my code as well, and we conformed 
our code to the later gcc interpretation or implementation. 
Not the end of the world, and not worth more than a passing 
observation once fixed.

> Part of this situation seems (to me) to be more of a blurring of backup
> -vs- RAID. Both are good and have a place. Either by itself is
> non-optimal.

properly /boot is _just not used_ once booting is done, 
except for kernel updates -- it can be wholly umounted if on a 
separate partition (as is a customary practice by some)

all of the above, my $0.02

-- Russ herrold