From owner-freebsd-questions Fri Jun 27 05:29:41 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id FAA19028 for questions-outgoing; Fri, 27 Jun 1997 05:29:41 -0700 (PDT) Received: from palrel1.hp.com (palrel1.hp.com [156.153.255.235]) by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id FAA18943; Fri, 27 Jun 1997 05:27:54 -0700 (PDT) Received: from postbox.india.hp.com (postbox.india.hp.com [15.10.45.1]) by palrel1.hp.com with ESMTP (8.7.5/8.7.3) id FAA23427; Fri, 27 Jun 1997 05:27:40 -0700 (PDT) Message-Id: <199706271227.FAA23427@palrel1.hp.com> Received: from localhost by postbox.india.hp.com with ESMTP (1.39.111.2/16.2) id AA123114416; Fri, 27 Jun 1997 17:56:56 +0530 To: Nik Clayton Cc: questions@freebsd.org, jfieber@freebsd.org Subject: Re: "make world" and re-making the kernel In-Reply-To: Your message of "Thu, 26 Jun 1997 10:13:56 +0100." <19970626101356.13640@blueberry.co.uk> Date: Fri, 27 Jun 1997 17:56:56 +0530 From: A Joseph Koshy Sender: owner-questions@freebsd.org X-Loop: FreeBSD.org Precedence: bulk >>>> "Nik Clayton" writes Hi Nik, I noticed that there was no response to your query from -questions ... perhaps you may have better luck with -hackers ... nc> [Talking about http://www-local/technical/freebsd/upgrading-from-src.html] jk> Basically I wanted to say that its recommended to rebuild/install the new jk> kernel before doing the reboot step in "8". nc> Is it? I was under the impression that it's better off to reboot the nc> system, so that all your 'make world' changes have definitely taken effect, nc> and then re-build the kernel. nc> Copied to -questions for informed comment. Heres my rationale for my statement: Rebooting with a new `world' but an old kernel can be dangerous, because many of the system utilities may break: (a) ps, ifconfig, sysctl etc. can fail to work properly, leaving the machine off the network or worse ... (b) worse, basic utilities like `mount' can fail leaving you unable to mount `/usr' or even `/' (this happened to me once when Lite2 was being merged in) ... (c) I don't use LKMs but I can envisage a scenario where an old kernel may crash with a newer LKM built during a make world. Having been bitten by (a) and (b) I now try to keep /kernel in sync with the world. The larger the gap between your current source base and the new `world', the greater your likely exposure to problems of this nature. Regards, Koshy