It would appear that on May 13, Gar Nelson did say:
> I'm not a trusting soul. I am running FC1 and Win4lin, and couldn't
> really do my job without either using a windows machine, or running win4lin.
>
> In my /etc/yum.conf I have added "exclude=kernel*" under the [main] block.
>
> Yum will no longer update kernels, which I didn't trust it to do by
> itself anyway, but will update everything else.
Thanks Gar,
I do tend to trust yum. I don't completely understand it however. But a
quick :! less /etc/yum.conf makes me pretty sure that this would be put
on it's own line anywhere in that section?
For the reasons described below, I don't think I'm going to use win4lin
until I can use yum to automate the kernel updates.
But one thing might change that. Your a win4lin user. Perhaps you
can explain something to me? If I understand the legalities of it I have
to feed it a (legal) copy of windows (such as my win98. Can it be
configured ( by a non-expert ) to just use my existing win98 installation
in place on hda1? (I'd back up win98 1st, so I wouldn't really care if I
lost native mode windows in the process.) But if it can use the current
* 5gig windows partition I shouldn't have to sacrifice nearly * 4gig of my
precious linux space to accommodate it. Though I suppose it could be
cloned into a virtual partition and then change the filesystem type of
hda1 to linux, and move the virtual windows directory tree to it...
Also, once the virtual win98 filesystem is in place, in it's own
mountable partition, could a 2nd copy of win4lin in the other fedora
filesystem be configured to use it so that regardless which fedora I
booted, I could still access it without rebooting???
If both of those are true, I might decide to manually do my kernel
and exclude the kernel from yum...
* :r!df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/hdc3 15377852 3914076 10682616 27% /
none 127772 0 127772 0% /dev/shm
/dev/hdc8 4806904 2392256 2170464 53% /the_j-home
/dev/hda1 5114684 3663524 1451160 72% /theWin98
Currently I do trust yum to update my kernel. Of course, since I'm using
lilo, it still takes manual intervention from me to actually implement
the use of the new kernel. I have a multi-boot pc and actually use the
lilo from my mdk9.1 (it still supports my beloved TEXT MODE MENU boot
screen.) But I configure lilo, and make boot floppies from each of my 4
linux systems... this makes kernel updates aggravating. But in case of
security fixes that I probably had my head in the sand too deeply to
know when it's really needed, I tend to do them all.
Actually Fedora is my very 1st attempt to keep any os up to date. The Yum
method is almost painless so it's worth a try... My strategy is to use
TWO of my four 14.9 gig linux partitions for fedora. The theory is that
I archive & clone my FC1 with tar and edit the new fstab and a few scripts
(I'm in the middle of doing that now) If FC2 wasn't so close, I'd update
one of them at a time. If after a week nothing seamed broke, I'd update
the other. and in case there were more things updated I'd start
predominantly using the clone fedora, then after a week I'd update the
main fedora switch back to it... (My personal mail folders and almost all
my personal data storage is actually in symlinked directories in a separate
partition, so switching the primary use from one fedora to the other is
kinda transparent. While allowing for potentially different ~/.* files {this
means I almost don't notice the difference even when I boot my mdk9.1})
When a stable core release come along, I upgrade ONE of them... 1st I
try the upgrade/update (whatever anaconda calls it) if it breaks then I
can wipe THAT fedora partition for a new install. And then strategically
copy my user data from the other... Anyway Once I have ALL the same
functionality with FC{new_stable} as with FC{previous_stable} I can
either repeat the process with the other or extract it from my tar
archive and reedit fstab etc... whichever was less work. Meaning if the
upgrade method worked I'd do it again, if I had to do NEW install I'd
probably use tare & edit...
Anyway once both are the same FC{stable_release} I'll again use the
alternating weekly update method described above, till the next core
release.
At least thats my plan. For the most part I'm confident that the people
who do the fedora (and freshrpm) packages are a *LOT* better than I am at
making it all work right. And this way I should always have an in-place
back up system should something get well and truly broken...
--
| --- ---
| Joe (theWordy) Philbrook <o> <o>
| J(tWdy)P ^
| <<jtwdyp@xxxxxxxx>> /---\ "bla bla bla..."
| \___/ "...and bla..."
At least I know my mouth is running, I just can't find the off button!
