On Thu, 2002-02-28 at 12:55, Craig S. wrote:
> I am preparing to upgrade my kernel. A friend of mine who does support
> for BRU Linux said that 2.4.18 was unstable for him and he upgraded to
> kernel 2.4.19 and is fine now. I have heard some similar input from a
> few other people. While discussing this in another forum I was told that
> the 2.4.19 kernel which is in the testing /v2.4/testing at
> ftp.kernel.org is just a patch. Now my friend's system reports the
> kernel as version 2.4.19 and as far as I can tell when I look at the
> version report it looks like a whole new kernel and not a patch of
> 2.4.18.
>
2.4.18 is the most stable 2.4 kernel so far, but it does have minor
problems (or not so minor if you are using it for NAS). The things I've
heard that are issues are:
- smbfs problems (random core dumps, can be nasty)
- won't build on Sparc64.
As far as I know (from reading the linux-kernel list) the current
released kernel is still 2.4.18, 2.4.19pre2 is available but is not a
final release (thus the EXTRAVERSION 'pre2').
If your friend does a 'uname -a' and does not see 'pre2' then something
is screwy.
> I would assume a patch just installs new code in a few specific areas of
> the kernel rather than updating the majority of the functions in the
> kernel. This seems as though it could be considered a new kernel though.
> My limitation of kernel structure I believe is my weak point here.
>
> I guess my question is, what is the definition of a patch vs a new
> kernel?
The patch will add the changes between 2.4.18 and 2.4.19pre2 (read the
changelog for info about what changed...)
If you do end up using 2.4.19 I'd recommend going with Alan Cox's tree
(currently 2.4.19pre2-ac2) as he tends to have more stable/reliable
kernels than mainstream pre-releases.
--
Blake Barnett (bdb) <blake.barnett@developonline.com>
Sr. Unix Administrator
DevelopOnline.com office: 480-377-6816
Learning is a skill, you get better at it with practice.
(text/plain)