FreeBSD has serious problems with focus, longevity, and lifecycle

[Mark Felder]

On Tue, 17 Jan 2012 12:59:44 -0600, Julian Elischer <julian at freebsd.org>  
wrote:

> On 1/17/12 9:36 AM, Damien Fleuriot wrote:
>
> having run -stable on production systems, the way to do it is:
>
> * follow -stable..
> * pick a time that IN RETROSPECT (from 1 month later) looks as though it  
> was good.
> * take a snapshot from that time and test it.
> * if it has problems MOVE FORWARD (not back) to the next candidate  
> snapshot time.
> * repeat until you have one that works for you
>

This is also the way I've had it explained to me. Note, I'm currently not  
running anything -STABLE in production right now simply because I don't  
have that need. I did for a while last year, but not now.

I'm deploying two ZFS SANs based on 9 early February. It might be on  
-RELEASE with manual backports of the gmultipath rewrite (required) and  
also I am considering the fixes for CDROM access (CAM stuff). I've seen  
several other things hit -STABLE right after the freeze ended early  
January which surprise me that they weren't included in -RELEASE and we  
didn't have another RC. I definitely see the frustration being expressed  
here, but I personally am comfortable running -STABLE. Many people are not  
and it is unfortunate that they will be left waiting until 9.1-RELEASE  
(probably late summer?). I hope a compromise will be found. I'm clearly in  
the minority that is OK with the current situation.

To be fair, it could be worse -- OpenBSD secretly wants you to run  
snapshots and CURRENT as the RELEASEs are mostly unmaintained outside of  
the most extreme security concerns. Even the packages are kept at the  
exact version of the time of release.

Well to be honest, Theo doesn't want me running OpenBSD at all. :-)


-Previously Flamed User