Date: Sat, 26 Sep 1998 19:19:41 -0700 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: Studded <Studded@dal.net> Cc: John Birrell <jb@cimlogic.com.au>, current@FreeBSD.ORG Subject: Re: Upgrade documentation (Was: Re: Make world error on -current elf) Message-ID: <1094.906862781@time.cdrom.com> In-Reply-To: Your message of "Sat, 26 Sep 1998 16:22:38 PDT." <360D773E.E0178B7F@dal.net>
next in thread | previous in thread | raw e-mail | index | archive | help
> I administer several systems that are 600 miles away, one that's 2,000. > Source upgrades are the only path for me, and I know there's a lot of > others in the same boat. There's no substitute for having someone on-site to handle a major upgrade. I also would not suggest upgrading a system that way, it being akin to doing brain surgery in the dark, no matter how creative the administrative situation was. Maybe with a point release upgrade I might try to get away with it, but certainly not for a major one like this. And, yes yes, I'm sure many people get away with doing all kinds of crazy stuff out there in the field, but hardly makes it recommended (or supported) procedure. :) > The problem is, within the next X months, 3.0 isn't going to be > -Current anymore, it's going to be the mainstream. IMO we need to start > planning for this change, oh, 6 months ago or so. :) 3.0.0-RELEASE has also been strongly discouraged for more than the last 6 months to any of our users who are halfway serious about reliability, and anyone who's been ignoring our warnings does so only at their own peril. Those who subscribe to the Walnut Creek CDROM distributions even got a letter in which they were essentially discouraged from buying 3.0 at this time unless they wanted it just for hobbiest/experimental purposes, and that should tell you something since folks like WC are in the business to make money and if they're telling you NOT to buy something, it's probably worth listening to them. People also ski in posted avalanche areas (and die) all the time, but that doesn't make them particularly smart. :) > If I can digress for a moment, this is actually a much bigger paradigm > shift than most people are realizing it seems. For years now, the 3.0 > branch (and I mean that specifically, as opposed to the more nebulous > idea of "-Current") has been the exclusive domain of the "elite" group > of hackers who could stomach its idiosyncracies. In a sense, it's become And our "dot-zero" releases are still that way. They've ALWAYS been that way, in fact, and we've never made any bones about it. We'll certainly be happy to see some of the "elite" folks (though I wouldn't use that word myself) jump on 3.0-RELEASE so that we can start getting some of the testing which will be required to turn 3.0-CURRENT into something that's eventually ISP material, but you can bet that ftp.cdrom.com won't be running 3.0 for awhile. When we're confident enough to switch our "flagship" machine to 3.0.x-RELEASE, then maybe the rest of our commercial customers should consider following suit. Until then, 2.2.8 is the recommended (and planned) upgrade path and that should hopefully take us through the end of the year. Come Q1 '99, we can start talking more seriously about these sorts of transitions being commonplace. > I've already seen some of the resentment, bad feelings, etc. associated > with this kind of shift appearing on the lists, and I expect more to Hmmm. I haven't seen much resentment coming from anyone but you. :-) Seriously, anyone getting "resentful" at this stage about the transition is simply trying to do the wrong thing at the wrong time. We have a release track for those who need strong shock-absorbers on their release and it's called 2.2-STABLE. I don't think people are taking that fact seriously enough and it's a damn shame, especially when they start blaming *us* for their own refusal to heed the warnings. - Jordan To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1094.906862781>