Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 03 Nov 2004 13:27:57 -0700
From:      secmgr <security@jim-liesl.org>
To:        aw1@stade.co.uk
Cc:        freebsd-stable@freebsd.org
Subject:   Re: freebsd 5.3 have any problem with vinum ?
Message-ID:  <41893F4D.6090702@jim-liesl.org>
In-Reply-To: <20041103031316.A95136@titus.hanley.stade.co.uk>
References:  <02f201c4ba91$f9f95db0$33017f80@psique> <1098725440.5103.4.camel@sp4.cs.ucdavis.edu> <417D604D.4090800@jim-liesl.org> <20041028153756.GB1195@eucla.lemis.com> <41856FAC.2020808@xena.ipaustralia.gov.au> <20041103031316.A95136@titus.hanley.stade.co.uk>

next in thread | previous in thread | raw e-mail | index | archive | help
Adrian Wontroba wrote:

>On Mon, Nov 01, 2004 at 10:05:16AM +1100, Carl Makin wrote:
>
>  
>
>>Do you want to yank it in 5 or 6-CURRENT?  There are a *lot* of people 
>>using vinum and yanking it in 5-STABLE would force us all to use the 5.3 
>>security branch until gvinum caught up.
>>    
>>
>
>>From my experiences today with setting up a very old machine[1] with
>5.3-RC2, I think it would be best to keep both until gvinum had caught
>up.  Vinum can do things which gvinum appears incapable of - such as
>initialising a RAID-5 plex.
>
>  
>
Just ran into this myself.  I had a perfectly happy raid 5 plex under 
5.3 RC1.  I upgrade to RC2, and the whole plex goes stale.  I deleted 
everything from the volume on down (except for the drives), and tried to 
recreate the vol/plex/sd's.  gvinum creates them, but they come back 
(like the undead) as stale and unusable (just like they were before).  
I'm finding commands documented (in help), but unimplemented 
(checkparity?  init?).  I hate to sound like a whiney baby, but WTF is 
going on?  It feels like vinum from 4.x has basicly been abandoned 
(short of crashes with no workaround), and gvinum ain't near ready for 
primetime.  We need a stable, working s/w raid solution (or admit that 
as of right now, FreeBSD doesn't have one).  At the very least, we need 
re and the authors to document what works, what doesn't, and what never 
will.  I'd happily help with docs, but right now it seems like product 
function is linked to /dev/random.  If gvinum isn't ready for release 
(my personal opinion after RC2), it needs to be pulled until whats 
documented works, and what works, is documented correctly.

</whiney_baby>
jim



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?41893F4D.6090702>