Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 26 Mar 2003 16:38:15 -0500
From:      taxman <taxman@acd.net>
To:        James Schmidt <James@JamesSchmidt.com>, questions@freebsd.org
Subject:   Re: Questions about upgrading from 4.6.2 to 4.7 (and beyond)
Message-ID:  <200303261638.15931.taxman@acd.net>
In-Reply-To: <5.1.1.6.0.20030326020015.00b29e20@207.254.222.218>
References:  <5.1.1.6.0.20030326020015.00b29e20@207.254.222.218>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wednesday 26 March 2003 03:19 am, James Schmidt wrote:
> I apologize in advance for my unfamiliarity with the workings of the
> FreeBSD world.  Out of the several hundred machines I administer I have
> only the one FreeBSD box, and it is considered a production server, so I
> don't get much chance to play around with it.
>
> At any rate, one of the tasks I'm going to be undertaking shortly is to
> migrate it from 4.6.2-RELEASE-p3  to 4.7-RELEASE (and soon after that, to
> 4.8) and I have several questions.  Actually, to be more precise, I would
> just like someone to verify that what I am doing is correct.
>
> First, the overall cvsup process - I want to make sure I am doing this
> right, someone please let me know if any of this is incorrect or
> inefficient.  Below is my cvsupfile:
>
> *default  host=cvsup13.FreeBSD.org
> *default  base=/usr
> *default  prefix=/usr
> *default  release=cvs tag=RELENG_4_7
> *default  delete use-rel-suffix
> *default  compress
> src-all

> # added manually... collection will be filtered through the refuse file
> ports-all tag=.

I put the ports tag into a separate supfile called ports-supfile, but you've 
got the right tag for inclusion in your supfile.

> I should be using the RELEASE branch, and not STABLE or CURRENT, correct ?

Yes, with a production machine

> The process I follow to cvsup is as follows:
>
> # cd /src/cvsup
> # cvsup -g -L 2 ./cvsupfile
>
> Once that is complete, the next steps are to build the kernel and the
> sources - and up until now, I've always compiled the kernel the
> old-fashioned way (config, make depend, make, make install) - with 4.7,
> should I begin using the new and improved method of kernel compiling, as
> follows ?

yes. you can also add some steps here for extra safety

backup /etc   this cannot be underestimated if you are not confident in your 
ability to recreate them from memory.

> # cd /usr/src
> # make buildworld

# mergemaster -p

> # make buildkernel KERNCONF=<kernelname>

Obviously replacing <kernelname> with the name of the kernel config file you 
have edited or GENERIC

> # make installkernel KERNCONF=<kernelname>
> <reboot>

drop to single user mode here with a 
shutdown now  from root.

> # make installworld

Then 
mergemaster 
Be careful here.  This will merge in the new scripts and config fles in /etc  
any that you have modified you will not want to automatically merge the new 
files in or you will lose your changes.
Some you'll want to merge by hand.

> I am sorry to be asking such basic questions.  I've read through the

Like Subhro said, no worries to ask questions as long as you've done your 
reading.  I'd suggest reading that handbook chapter again, and also reading 
/usr/src/UPDATING as that file has the final say.

> pertinent sections of the online FreeBSD manual  and I *think* what I am
> doing is correct, but I want to make absolutely sure before I reboot the
> machine and end up with a kernel panic, umountable root or something
> equally as horrific.  If anyone sees any glaring omissions or any other
> problems with what I'm doing, please let me know.  Again, many thanks.

You're welcome, good luck

Tim



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200303261638.15931.taxman>