Subject: Re: Why was vinum removed?
To: Mathieu Boespflug <>
From: Greg 'groggy' Lehey <>
List: current-users
Date: 05/02/2006 13:28:43
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Monday,  1 May 2006 at 17:46:09 +0100, Mathieu Boespflug wrote:
> Hi,
> I was looking at some recent CVS commits and noticed that vinum was
> removed. After looking through the archives I haven't found any
> discussion of this and why this was done. Is there now any form of
> volume management under netbsd ala linux LVM or freebsd GEODE +
> gvinum?

The background is that Vinum was broken, and I didn't have any time to
look at it.  I was approached by core in early February and asked if I
would object to its removal.  I replied:

On Friday, 10 February 2006 at 10:45:25 +1030, Greg 'groggy' Lehey wrote:
>> The core team has been looking at some of the lesser-used subsystems
>> in NetBSD, and we've been looking long and hard at vinum. There
>> are a number of open PRs outstanding - notably 29129, 23462, 25547
>> and 32257, and I think that it's fair to say that vinum, whilst
>> providing useful functionality, is not in a state that we could
>> recommend to people for use in a production environment.
>> The problems are exacerbated in that vinum is not being maintained,
>> and fixes are not brought in from other archives or from the development
>> branch. (PR 23462 has been open since November 16th 2003, PR 25547
>> since May 12 2004).
>> Because of these problems, we'd like to remove vinum from NetBSD's src
>> tree. We can do it ourselves, if you'd like us to, or you can remove
>> it yourself.
> I can't blame you.  I just haven't had time to maintain Vinum on any
> platform.  If your word is final, go ahead and remove it.
> Before we axe it, though, would it not be worthwhile looking round for
> somebody else to maintain it?  As you say, it's a useful subsystem to
> have.  If we can find somebody else to share the load, I should be
> able to help get it up to speed.

I didn't get any reply to that message.

> If there is no LVM, I would be interested in participating in as
> much of implementation effort as possible.

In that case, I think that it would be well worthwhile bringing it

The biggest issue that I know about are (apart from my own lack of
time) is that incorrect major device number may mean that the thing
doesn't work at all.  I'm not sure of the current status of this.  If
anybody knows of other problems, please let me know and we can discuss
a resolution.

See complete headers for address and phone numbers.

Content-Type: application/pgp-signature
Content-Disposition: inline

Version: GnuPG v1.4.1 (FreeBSD)