Possibility for FreeBSD 4.11 Extended Support

[Charles Sprickman]

On Thu, 21 Dec 2006, Bill Moran wrote:

> In response to Heinrich Rebehn <rebehn at ant.uni-bremen.de>:
>
>> Colin Percival wrote:
>>> John Smith wrote:
>>>> Support for FreeBSD 4.11 is going to end sometime in late January.
>>>> Originally, FreeBSD 6.2 was supposed to be released back in October.  This
>>>> would have given everyone about 3 months to stress test everything and
>>>> migrate all their boxes from 4.11 direct to 6.2.
>>>
>>> You've had three months to stress test 6.2-BETA1, 6.2-BETA2, 6.2-BETA3, and
>>> 6.2-RC1.  We release these for a reason, you know.
>>>
>>>> Now it is near the end of
>>>> December, and FreeBSD 6.2 RC2 has yet to be seen anywhere.  Chances are that
>>>> FreeBSD 6.2 Release will come out earliest mid-January.  This does not give
>>>> much time for people to migrate to the newest FreeBSD release.  I think it
>>>> would be fair if support is extended for a few more months especially since
>>>> 6.2 is so late in coming.
>>>
>>> Your opinion has been noted.
>>>
>>> Colin Percival
>>
>> I have to second the OP's opinion. :-)
>> I think it is important to be able to stress test the *final* release
>> before installing on production machines. There is little use in stress
>> testing BETAs and then install a broken RELEASE.
>> This happened with 6.1-RELEASE where the nfs server was suddenly
>> unusable on amd64.
>
> There is something about these "please continue to support 4.x"
> discussions that confuses me.

Personally, I understand it, but my perspective may be skewed.

> The general argument has been that 4.11 support should continue because
> 6.2 is not at release status yet.

If I were to complain about 4.11 going away (and I'm not), this would be 
my argument:

-5.x was never really for production use, in the same way 3.x never was. 
It was a release made to introduce new features and to beta test what will 
become a good 6.x release.  In my mind, I always skip over 5.x.  I would 
not shed a tear if support for 5.x was dropped before 4.11.

-the 4.x branch was the most stable thing since 2.2.x, so many people are 
hanging on to it for dear life, much as in the windows world you'll still 
find people that prefer the (relative) stability of something like W2K 
over XP or Vista.  It is a *compliment* to everyone that put all the 
effort into making the 4.x branch as good as it was that people want to 
keep using this functional and stable software.

-many people run a ton of machines and are not doing any hardware swaps 
anytime soon.  4.11 runs well on there, and doing a full reinstall on 
dozens, hundreds or thousands of hosts might be more than they care to do 
*right now*.  Again, a testament to the stability and quality of 4.11.

-upgrades from 4.11 to 6.2 are not simple, and not doable without a fairly 
significant amount of downtime.  Everywhere there are folks with a handful 
of boxes that shouldn't be a single point of failure, but are.  Worse, 
some people have a mix of unique boxes where their first test of 6.x is 
going to be their only test of 6.x on that specific piece of hardware.

-there certainly are plenty of new features and conveniences in 5/6, but 
for a 3 or 4 year old box that's happily humming along, new hardware 
support is not paramount, nor are things like the vastly improved wireless 
support.  In any sort of large server farm there are likely homegrown 
solutions in place to augment 4.11 to the point where the lack of 
/etc/rc.d or other little convenience pieces just aren't compelling enough 
to start over.

> Are the people making this argument unaware that 6.1 and 5.5 have been
> at release status for quite some time, and thus have been providing
> ample opportunity to upgrade for some time now?  Or has this topic
> simply degraded to Troll bait?

Again, I think 5.x is probably the least used version of FreeBSD in 
history.  As for 6.1, using a .1 release of something in production is 
gambling (not a knock on FreeBSD, I'd apply that to anything).

People are just voicing their opinion.  This is not a democracy, but that 
also does not preclude the userbase from expressing their views on the 
matter.  If this were a democracy and this was a vote, I'd vote for 
extending 4.11 support until 6.3 comes out and dropping all support for 
5.x tomorrow. :)

FWIW, I have about 1/4 of the production boxes I manage up to 6.1 or 
6.2-RC1 (mostly throwaway/redundant stuff like spam scanning).  The rest 
are still at 4.11.  I do look forward to the bonuses of moving to 6.2 or 
6.3 on the rest; my short list of new stuff that would make my life 
easier: pf, the new rc stuff, jail improvements, support for more GigE 
interfaces, mysql almost working right/threads.

Charles

> -- 
> Bill Moran
> Collaborative Fusion Inc.
> _______________________________________________
> freebsd-stable at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-stable
> To unsubscribe, send any mail to "freebsd-stable-unsubscribe at freebsd.org"
>