raidframe in -CURRENT

[Scott Long]

Darcy Buskermolen wrote:
> On Friday 15 August 2003 03:07, Andrew P. Lentvorski, Jr. wrote:
> 
>>I'm not going to speak for Scott Long, but I actually looked into doing
>>something about RAIDFrame.
>>
>>The problem is that GEOM provides a significant superset of the same
>>functionality as RAIDFrame.  GEOM virtualizes different storage layers in
>>much the same way as RAIDFrame did (but with more flexibility).
>>
>>The problem is that Poul-Henning Kamp hasn't gotten around to implementing
>>the RAID virtualized modules yet.
>>
>>Consequently, FreeBSD is currently in a state where the GEOM equivalent to
>>RAIDFrame doesn't yet work, but doing the work to fix RAIDFrame is just as
>>much work as doing it in GEOM.  :(
>>
>>-a
> 
> 
> Scott, Poul-Henning, can either of you provide any insight as to when a viable 
> software RAID archetecure will be available in -CURRENT, I have tried vinum, 
> however its' inability to handel hotspare has turned it into a non option for 
> me at this time.
> 
> 

It's a matter of time, really.  It will probably take about 500 hours to
convert RAIDframe to GEOM and then make it robust.  If I was doing it as
a day job, that equates to 3 months.  Trying to do work on it for an
hour or two each evening really isn't possible.

I share your concern about the state of software raid in FreeBSD.  When
I started RAIDFrame I was in a position to devote full-time hours to it.
Now, I'm not, and I'm at a loss of what to do about it.

Scott