Date: Tue, 12 Sep 2000 00:41:55 -0700 From: Kent Stewart <kstewart@urx.com> To: Mike Nowlin <mike@argos.org> Cc: cjclark@alum.mit.edu, The Hermit Hacker <scrappy@hub.org>, Francisco Reyes <fran@reyes.somos.net>, Gregory Sutter <gsutter@zer0.org>, "freebsd-stable@FreeBSD.ORG" <freebsd-stable@FreeBSD.ORG> Subject: Re: 4.1-STABLE fails to 'buildkernel'? Message-ID: <39BDDE43.77FE9224@urx.com> References: <Pine.LNX.4.21.0009120256410.5541-100000@jason.argos.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Mike Nowlin wrote: > > > > What SUCKS is when you manage to get through a make-world without an > > > error, but the running system is broken. /That/ is something to whine > > > about. (And you just want to slug that first one that says, "Just use > > > the backups you made before the make-world." I do backup first, but > > > restoring is still a non-trivial pain in the *ss. And if you didn't > > > backup, especially a production system, you asked for it, bub.) > > > > This is all part of the reason why I do a buildworld, > > build[install]kernel, and finally do my installworld. When something > > like this happens, my kernel.old and my user land are still in sync. > > > > I kept my mouth shut at the time, since I knew it would be fixed in a > couple of days and the machine I was upgrading wasn't anything really > important, but the problem a couple weeks ago where telnetting into a > system or starting X on it caused it to blow up didn't exactly strike me > as "stable". kernel.old didn't work since the userland had changed enough > to make ps, top, etc. to stop working -- un-cvsup it and rebuild... > > This problem shows up more often when you update sources weekly or > so. Daily doesn't seem to be a big problem - kernel.old is usually > good; going from 4.0-RELEASE to 4.1-STABLE can cause headaches. I have 3 machines that are running builds within 2 or 3 weeks of each other. The two that can go down are built much more often. When everything works and there are changes I want, I cvsup and build my server. The other two can be down for a short while but the server isn't upgraded unless the other two are working. Things change and so it can't be too far behind. I will upgrade one of them if it appears something is broken. I haven't done it right now because I have some seti wu's that are getting too old. > > Just goes to show you that completely trusting > [build|install][world|kernel] and friends isn't completely > goof-proof. Creating a program that compiles flawlessly but still doesn't > work doesn't require much cranial activity. That is very true but you don't get hit by errors like that as frequent as you are by builds that won't finish. The changes are supposed to be tested before they are MFC'ed to stable from current. The telnet/ping problem existed in current and those changes probably shouldn't have been MFC'ed until they were fixed. With the common breakage, if you haven't installed anything, you haven't damaged anything and can wait for the "Maytag repairmen" to fix things :). Kent > > (BTW: Thanks, guys, for fixing that problem! Got my shiny new Toshiba > 4360ZDVD laptop running -STABLE now - X, sound, pcmcia, > everything! Although the S3 Savage-IX patch for XFree-3.3.6 has some > problems (with work-arounds), it's kicking along nicely!) > > - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - > > Understated/funny man-page sentence of the current time period: > > >From route(4) on FreeBSD-3.4, DESCRIPTION section: > "FreeBSD provides some packet routing facilities." > ...duh....... > > Mike Nowlin, N8NVW mike@argos.org http://www.viewsnet.com -- Kent Stewart Richland, WA mailto:kbstew99@hotmail.com http://kstewart.urx.com/kstewart/index.html FreeBSD News http://daily.daemonnews.org/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?39BDDE43.77FE9224>